Обсуждение: Server unreliability
It is my opinion that we have to make major changes in the way we provide hosting for our servers. There are several problems: o Location of servers The location of our servers in Panama is a problem. They are too far for any PostgreSQL maintainers to access. Changing hardware or diagnosing problems has been too hard. I have had like 2 days of downtime on my home machine in the past 12 years. We have had more than 2 days of downtime in the past 6 months. My wife would not accept such a reliability level. o FreeBSD The use of FreeBSD jails can cause servers to take +8 hours to fsck on a server crash or power failure. Again, I would never accept such problems on my home server so it is hard to fathom how a project with thousands of users can accept that. Either we need to find a fix, stop using jails, or get another operating system, but continuing to use a setup with a known problem is just asking for trouble. o Web site We have been talking about a new web page layout for years at this point. I almost don't care if they just put a dancing bear up on the web site. Let's do something! o Archives The archives situation is a continual problem. Again, maybe a dancing bear can help. :-) Basically, with no money and no one offering servers, I don't see a good solution to any of these problems, but I think we need to recognize these are problems and that we will continue to suffer until they are addressed. Are there any proposals, no matter how radical, to correct these? -- Bruce Momjian | http://candle.pha.pa.us pgman@candle.pha.pa.us | (610) 359-1001 + If your life is a hard drive, | 13 Roberts Road + Christ can be your backup. | Newtown Square, Pennsylvania 19073
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 Hi, On Wed, 29 Sep 2004, Bruce Momjian wrote: > o Location of servers > > The location of our servers in Panama is a problem. They are too far > for any PostgreSQL maintainers to access. Changing hardware or > diagnosing problems has been too hard. I have had like 2 days of > downtime on my home machine in the past 12 years. We have had more than > 2 days of downtime in the past 6 months. My wife would not accept such > a reliability level. I think mirroring of the web & FTP sites helped hub.org too much, but I still agree with you. Do you have any suggestions for that? Who would/could support/donate such a bandwidth? > o FreeBSD > > The use of FreeBSD jails can cause servers to take +8 hours to fsck on a > server crash or power failure. Again, I would never accept such > problems on my home server so it is hard to fathom how a project with > thousands of users can accept that. Either we need to find a fix, stop > using jails, or get another operating system, but continuing to use a > setup with a known problem is just asking for trouble. I hope we'll not begin a OS/distro war! (O.K, it'd be good if Red Hat contributes RHEL to PostgreSQL project ;) Anyone from Red Hat in here, or in -core? ;) ) > o Web site > > We have been talking about a new web page layout for years at this > point. I almost don't care if they just put a dancing bear up on the > web site. Let's do something! Like this one? : http://www.jensummerfield.supanet.com/bearaniy.gif Heh heh. Well, web team (mostly Alexey) is working on it. We have a fabulous new design (you know...). I think it'll be ready before 8.0.0. See the -www archives for that (Archives? Umm maybe Google...) > o Archives > > The archives situation is a continual problem. Again, maybe a dancing > bear can help. :-) :-) A Google search box could be better sometimes... Regards, - -- Devrim GUNDUZ devrim~gunduz.org devrim.gunduz~linux.org.tr http://www.tdmsoft.com http://www.gunduz.org -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.2.1 (GNU/Linux) iD8DBQFBWvA6tl86P3SPfQ4RAkzLAKCVF+7lDhNbyWW00x9yXpXVdxew/gCeOp1C Mf0/wlCZiX5ERu4e7Ef2xPc= =NGbG -----END PGP SIGNATURE-----
I think mirroring of the web & FTP sites helped hub.org too much, but I still agree with you. Do you have any suggestions for that? Who would/could support/donate such a bandwidth?I'm always willing to donate such things. I've got about 80mbit unused right now.
I hope we'll not begin a OS/distro war! (O.K, it'd be good if Red Hat contributes RHEL to PostgreSQL project ;) Anyone from Red Hat in here, or in -core? ;) )
Heh, I use Gentoo if it makes any difference.
Well, web team (mostly Alexey) is working on it. We have a fabulous new design (you know...). I think it'll be ready before 8.0.0. See the -www archives for that (Archives? Umm maybe Google...)
Hopefully the logo banner gets updated with my change or the author can clean it up a bit... the current one is too fuzzy, the glassy logo is nice but didnt resize well. I can make one sized correctly if wanted.
Gavin
On Wed, 29 Sep 2004, Bruce Momjian wrote: > It is my opinion that we have to make major changes in the way we > provide hosting for our servers. There are several problems: > > o Location of servers > > The location of our servers in Panama is a problem. They are too far > for any PostgreSQL maintainers to access. Changing hardware or > diagnosing problems has been too hard. I have had like 2 days of > downtime on my home machine in the past 12 years. We have had more than > 2 days of downtime in the past 6 months. My wife would not accept such > a reliability level. This is currently being worked on ... we are looking at various remote management solutions so that we don't have to deal with waiting for a technician to get 'on the scene' ... > o FreeBSD > > The use of FreeBSD jails can cause servers to take +8 hours to fsck on a > server crash or power failure. Again, I would never accept such > problems on my home server so it is hard to fathom how a project with > thousands of users can accept that. Either we need to find a fix, stop > using jails, or get another operating system, but continuing to use a > setup with a known problem is just asking for trouble. Actually, again, this one is being addressed ... there is a solution in the pipeline to fix the cause of the 8+ hour fsck, but, since it is a fix to fsck itself, it hasn't been put into the mainstream code yet, due to *obvious* testing reasons ... We've also added in hot failover as an option ... I've posted to -www asking about putting www.postgresql.org onto it, but so far, the only responses back have been along the lines of 'how are you doing it?' ... The "risk" is that its not real time replication between the live and failover server ... on our high performance servers, the 'delay' is about 5 minutes ... ... now, knowing that, if you feel comfortable with me putting this onto the mailing lists/cvsroot as well, knowing that there is a possibility of something being written 'in the gap' before failover, I'll do that VM also ... note that altho the replication has a gap, the heartbeat process runs every minute ... as soon as it can't ping anymore, it fails over ... > o Web site > > We have been talking about a new web page layout for years at this > point. I almost don't care if they just put a dancing bear up on the > web site. Let's do something! What's wrong with the existing one? Have you designed the dancing bear you'd like us to put up in place of what we have now? > o Archives > > The archives situation is a continual problem. Again, maybe a dancing > bear can help. :-) What is wrong with it now? I'm cleaning up the code itself, but that is due to it being a mess right now, not due to any problems reported to me, removed one of the banner ads so that loading is a bit faster, and John has done, I think, a fantastic job on the search engine itself, including sending me changes for the archives themselves so that the 'time searches' should now work properly ... So, do you have something specific you'd like to point out to us that we've overlooked and haven't fixed yet? > Basically, with no money and no one offering servers So far, I've had one person donate $10 ... in order to put a dedicated server onto the network, I'd need alot more of those ... that would pretty much eliminate your second point about the fsck's, since its only our *loaded* servers, that we have that problem with ... but, as I said, the fsck issue is being addressed as well ... ---- Marc G. Fournier Hub.Org Networking Services (http://www.hub.org) Email: scrappy@hub.org Yahoo!: yscrappy ICQ: 7615664
On Wed, 29 Sep 2004, Gavin M. Roy wrote: > /I think mirroring of the web & FTP sites helped hub.org too much, > but I still agree with you. Do you have any suggestions for that? > Who would/could support/donate such a bandwidth? > > / > > I'm always willing to donate such things. I've got about 80mbit unused > right now./ a few things to note ... the only 'critical' thing right now that can't be mirrored is the cvsroot/mailing lists ... web and ftp are both well mirrored, and, in fact, ftp.postgresql.org is on a seperate network/server that I have ... archives was also moved there, and over the next few days, will be pointed to one of commandprompt's web sites ... in fact, bittorrent, ftp and archives are all on a US based server ... what I'd like to do is setup www.postgresql.org to point, via dns round robin, to the www.us.* mirrors, so that hits to it are more load balanced, but that requires some major changes to the mirrors themselves, and few of them are configured to recognize a virtual host of www.postgresql.org right now, as I found out not too long ago :( ---- Marc G. Fournier Hub.Org Networking Services (http://www.hub.org) Email: scrappy@hub.org Yahoo!: yscrappy ICQ: 7615664
On Wed, 29 Sep 2004, Robert Bernier wrote: > Could somebody explain to me how panama came to be chosen. Those reasons > may have changed since then. Client requirements ... and, quite frankly, the issues that Bruce is bringing up are in the works of being fixed by getting some remote management hardware so that I can deal with 99% of the problems from here ... > I'm very comfortable using Debian and would be willing to remote admin > using that. As for sticking with FreeBSD its the better choice if > there's somebody up to speed with security. The best resource person I > know is OReilly's FreeBSD columnist Dru, dlavigne6.sympatico.ca. She > might be convinced to "advise" what to do. the problem isn't a security issue ... the problem is that the unionfs creates these 'zero length directories' all over the place in the normal course of operations. fsck was never coded to handle them properly, so when it does do the fsck, it goes through a removes them all (not harmful, just time consuming) ... one of the developers was recently able to come up with a patch to correct the behaviour, but is rightfully nervous of putting it into the source tree, so is doing some more testing on it first ... As I also mentioned in the other thread, we've setup a hot failover option that I'd like to put on the VMs, but haven't had any good feedback on the -www lists about doing this ... for www.*, I don't believe there is any risk with doign it, but for stuff like pgfoundry/gborg and mailing lists, those VMs all have a cvsroot, so am a bit afraid to have it hotfailover and potentially lose someone's commit ... if ppl feel that is an acceptable risk to avoid the lengthy downtimes, I can enable it on all of the *.postgresql.org VMs right now ... but I need some feedback on that first ... > How does the archives work i.e. what's running it? I hate to state the > obvious but can we have an itemized list as to what's wrong with them. > I've got my own ideas but I'd like to be on the same page as everybody > else. I'd love to see a list as well ... I thought that John/Dave and I had addressed all of the outstandings, but obviously Bruce doesn't feel the same way ... or, of course, he could be doing like the last time and rambling off past problems that have since been fixed :( ---- Marc G. Fournier Hub.Org Networking Services (http://www.hub.org) Email: scrappy@hub.org Yahoo!: yscrappy ICQ: 7615664
On Wed, 29 Sep 2004, Darcy Buskermolen wrote: > The over use of jails I think is complicating things, what are the odds > that someone could find it in their heart to raid the old hardware pile > in their serverback room, and come up with a PIII 800MHz + system with a > gig of ram that they could donate to the cause, and we move some of the > more critical services (www/ftp/anoncvs) to that box without jails. jails are not the problem ... the problem is the file system we are using to share applications ... but, yes, if we could put a dedicated box in place, we could very easily eliminate the use of the file system ... we have never had a jail related crash on those servers, they have always come down to the file system itself ... > One thing I know we can do here to help at least the availability of the > website itself, is to use a round-robin DNS based setup. I know Marc > adjusted DNS records to point the main website at www3.ca a while back > during a time that www was stuck in fsck hell. To do this, and I'd love to, we'd need to have *at least* all of the www*.us.postgresql.org mirrors setup a ServerAlias for www.postgresql.org ... not sure what would be involved in getting that done though, but the DNS side wouldn't take much ... > What about combining the efforts od pgsql.ru and command prompt to do > this, or maybe even monarc or google groups ? Until someone can *list* the problems, how can a solution be suggested? John and I know of no issues at this time ... the last one was the 'time period' searches, which we believe we've just fixed ... > Ok i know how much we all hate banner adds and the like, but what about > using Google ads especially on the archives to help generate income to > fund these requirements? Huh? Google ads? I know what you mean, just how do we make use of them? ---- Marc G. Fournier Hub.Org Networking Services (http://www.hub.org) Email: scrappy@hub.org Yahoo!: yscrappy ICQ: 7615664
Bruce Momjian wrote: > Basically, with no money and no one offering servers, I don't see a > good solution to any of these problems, but I think we need to > recognize these are problems and that we will continue to suffer > until they are addressed. I'm sure people would offer money, servers, and pipes if there were some public accounting and governance. Right now, we don't really know - What services are being run? - How many resources does that take? - Who provides those resources? - Who administers these services? Take a look at this: <http://db.debian.org/machines.cgi>. This is the sort of thing we need before we can continue planning our system resources. -- Peter Eisentraut http://developer.postgresql.org/~petere/
On Wed, 29 Sep 2004, Mitch Pirtle wrote: > Hi Bruce, we met in Manhattan for your NYPHP presentation - I was the > lone PostgreSQL cheerleader in the back ;-) > > You ask for proposals, no matter how radical, so here goes: > > o Location of servers > > Pending bandwidth requirements, I am willing to sponsor hosting of > these sites, possibly on dedicated hardware. But I need some ideas of > bandwidth usage so I can figure out which strategies make the most > sense financially. BTW be glad your servers are not in Florida, I > have one client that is in a significant amount of pain right now. The server location is not changing, but thank you for your offer ... > o FreeBSD > > I'm agnostic in this regard, but have experienced that Linux distros > seem to be the easiest to maintain (Debian, Fedora). Again, operating system isn't changing, but thank you for your offer ... Both of the above issues that Bruce pointed out are being worked on/addressed, just not as fast as Bruce would like to see happen ... > o Web site > > I am a core developer on the Mambo project (www.mamboserver.com), and > joined the core team primarily to implement database abstraction > (currently tied to MySQL gasp!). Mambo has grown quite famous for > being easy to use, has lightweight requirements, and has a huge 3rd > party community that provides all kinds of extra goodies (forums, i18n > support, commerce, you name it). I would happily volunteer the > efforts to migrate the existing site to Mambo to make it easier for > the core gang to keep the site current. I would also consider having > PostgreSQL.org running Mambo being excellent incentive to prioritize > the implementation of ADOdb ;-) There is currently a re-write of the web site being worked on ... if you wish to aid in that, please feel free to subscribe to pgsql-www, where the discussions on this are taking place. Alexey(sp?) has spent a good deal of time working on/cleaning up the backend code itself, and there has been one site redesign proposal that I *believe* has been accepted already for how the new site will look ... the fun part is trying to find man hours to implement it :( ---- Marc G. Fournier Hub.Org Networking Services (http://www.hub.org) Email: scrappy@hub.org Yahoo!: yscrappy ICQ: 7615664
On Wed, 29 Sep 2004, Peter Eisentraut wrote: > Bruce Momjian wrote: >> Basically, with no money and no one offering servers, I don't see a >> good solution to any of these problems, but I think we need to >> recognize these are problems and that we will continue to suffer >> until they are addressed. > > I'm sure people would offer money, servers, and pipes if there were some > public accounting and governance. Right now, we don't really know > > - What services are being run? > - How many resources does that take? > - Who provides those resources? > - Who administers these services? > > Take a look at this: <http://db.debian.org/machines.cgi>. This is the > sort of thing we need before we can continue planning our system > resources. If someone wishes to create the form/interface, I'll fill in the blanks ... right off the top of my head, we have the following VMs running: svr1.postgresql.org - mailing lists, developer.postgresql.org, cvsroot - host server: neptune, Dual Xeon www.postgresql.org - techdocs, advocacy, www - host server: venus, Dual Athlon pgfoundry.org - host server: jupiter, Dual-PIII gborg.postgresql.org - host server: venus, Dual Athlon archives.postgresql.org - mailing list archives - moving to one of commandprompt.com's servers ftp.postgresql.org - also bittorrent and anoncvs - host server: saturn, leased computer pugs.postgresql.org - host server: venus, Dual Athlon search.postgresql.org - hosted on John's server Did I miss any? ---- Marc G. Fournier Hub.Org Networking Services (http://www.hub.org) Email: scrappy@hub.org Yahoo!: yscrappy ICQ: 7615664
On Wed, 29 Sep 2004, Darcy Buskermolen wrote: > On September 29, 2004 11:09 am, Marc G. Fournier wrote: >> On Wed, 29 Sep 2004, Darcy Buskermolen wrote: > ----8<----[snip] > > jails are not the problem ... the problem is the file system we are using >> to share applications ... but, yes, if we could put a dedicated box in >> place, we could very easily eliminate the use of the file system ... we >> have never had a jail related crash on those servers, they have always >> come down to the file system itself ... > > Poor use of terminlogy on my part, but the net effect is the same, the FS is > being used to facilitate the jail, take away the jail, take away the need of > the stacked FS. Actuallym, not true ... the FS is used to reduce the overall amount of disk space required, since ~90% of the jail is redundant ... when I can get a dedicated server in place, and only have the 10 or so VMs running on it, then disk space is no longer an issue, so I can easily get rid of the FS component ... that would get rid of the crashes ... > *nod* In the interim there is nothing stoping getting the process > rolling with those sites that are known to work, as well as contacting > the mirror admins. of the unknown sites.. I think you are the only 'known to work site' ... so, if I setup a round robin between the Panama and Canadian location for www.postgresql.org, would you be okay with the extra load? :) If so, I'll get to work on the coding changing required for the DNS ... > Specificly the program out lines at: https://www.google.com/adsense/ I > know of a few reasonably high traffic sites that are generating about > 5k/mth income from this program. 'k, will have to read through, but first scan, would the replace our current search system, or would the 'Google search box' be on the right of the screen, or ... ? Without me reading it all the way through, do you already know? ---- Marc G. Fournier Hub.Org Networking Services (http://www.hub.org) Email: scrappy@hub.org Yahoo!: yscrappy ICQ: 7615664
Bruce, > It is my opinion that we have to make major changes in the way we > provide hosting for our servers. There are several problems: You're not the only one to have noticed this, believe it or not. ;-) > The location of our servers in Panama is a problem. They are too far > for any PostgreSQL maintainers to access. Changing hardware or > diagnosing problems has been too hard. I have had like 2 days of > downtime on my home machine in the past 12 years. We have had more than > 2 days of downtime in the past 6 months. My wife would not accept such > a reliability level. To be fair, PostgreSQL uses an enormous bandwidth and traffic that gives our systems very little tolerance for anominalies. Not that having the stuff in Panama doesn't slow down fixes, but hosting postgresql.org is a bit more demanding than hosting, say, your or my personal web site. > The use of FreeBSD jails can cause servers to take +8 hours to fsck on a > server crash or power failure. Again, I would never accept such > problems on my home server so it is hard to fathom how a project with > thousands of users can accept that. Either we need to find a fix, stop > using jails, or get another operating system, but continuing to use a > setup with a known problem is just asking for trouble. I agree that when we have more servers dedicated to PostgreSQL, we should use a different approach than the jails. Marc, what are the advantages of the jails on a dedicated server, since we are so familiar with their drawbacks? > We have been talking about a new web page layout for years at this > point. I almost don't care if they just put a dancing bear up on the > web site. Let's do something! I'd suggest finding some money to sponsor somebody full-time for a month to accomplish the migration to the new web structure, which is more-or-less finished. That's what's holding us up right now AFAIK. > The archives situation is a continual problem. Again, maybe a dancing > bear can help. :-) But being solved. The archives have been copied to CommandPrompt, and the various archive search tools have been distributed. Once we can change the web interface, there should always be one tool and/or version available regardless of individual server/hosting failures. Further, SF.net has offered to host our mailing lists, which is tempting because of their direct mailing-list-to-database tool, as well as their more distributed hosting structure which prevents DNS-based failures. I expect to find out the terms for this next week, and then discuss it on WWW. > Basically, with no money and no one offering servers, I don't see a good > solution to any of these problems, but I think we need to recognize > these are problems and that we will continue to suffer until they are > addressed. > > Are there any proposals, no matter how radical, to correct these? The big obstacle in moving anything is bandwidth. While any number of companies will offer to host stuff for us, we can't match the amount of bandwidth we're currently using in Panama -- and hosting donors won't support anywhere near that level (which runs to scores of GB per month). So while we can move individual, less-crucial components, the mailing lists and the main WWW require either Hub.org's generous bandwitdh or other hosting that we "own". I have actually been working on raising money for additional servers. The obstacles which have made this a year-long process are more complicated than I want to go over right now, but there are people thinking about it. -- Josh Berkus Aglio Database Solutions San Francisco
Marc, Darcy, Bruce, Peter, et. al. > Poor use of terminlogy on my part, but the net effect is the same, the FS > is being used to facilitate the jail, take away the jail, take away the > need of the stacked FS. The other BSD experts I've talked to all agree that the Jail setup on the current servers is a severe performance drain. I think we really need to look at not using jails for the the dedicated PostgreSQL server(s), when we have them. > One thing I know we can do here to help at least the availability of the > website itself, is to use a round-robin DNS based setup. FWIW, some folks at SourceForge have offered their help in configuring this. I'm currently waiting for Adi to get back from vacation to proceed. > I'm sure people would offer money, servers, and pipes if there were some > public accounting and governance. Right now, we don't really know > > - What services are being run? > - How many resources does that take? > - Who provides those resources? > - Who administers these services? This is why I'd like to do some of this stuff -- particularly fundraising -- through the Foundation. However, as you know Peter, there have been some delays in that regard. -- Josh Berkus Aglio Database Solutions San Francisco
On Wed, 29 Sep 2004, Josh Berkus wrote: > I agree that when we have more servers dedicated to PostgreSQL, we > should use a different approach than the jails. Marc, what are the > advantages of the jails on a dedicated server, since we are so familiar > with their drawbacks? 'k, just to re-emphasize ... the jails are not, and have never been, the problem ... when I can put a dedicated server in, though, the thing that will fix the problem is getting rid of the unionfs that we're using to converse disk space ... as to what advantages the jail's provide ... mainly, security ... we can easily provide root access to pugs.postgresql.org, as an example, so that Fred (please tell me I remembered right?) can do whatever work he wants, without giving him full root on the developer jail ... same with all of the other jails ... Chris has full root on gborg, so that he doesn't have to ask someone else to do something when he notices a problem (ie. restart mailman) ... The other advantage ... seperate process space ... so, as an example, if someone wanted to run a different server then apache on port 80, they could ... > But being solved. The archives have been copied to CommandPrompt, and > the various archive search tools have been distributed. Once we can > change the web interface, there should always be one tool and/or version > available regardless of individual server/hosting failures. Joshua and I should have this switched over by end of day today ... I just did a bunch of "clean ups" to fix some of the issues with the search form, but to also balance things out a bit more ... also removed the right banner ad, so that page loading is a bit faster ... and I think we have the 'Last-modified' stuff finally working properly ... > The big obstacle in moving anything is bandwidth. While any number of > companies will offer to host stuff for us, we can't match the amount of > bandwidth we're currently using in Panama -- and hosting donors won't support > anywhere near that level (which runs to scores of GB per month). So while we > can move individual, less-crucial components And this is somethign that we have been working on over the past several months ... search has moved to John's location/servers, archives has done one move so far, and is going to be moving to CPs location, etc ... And, now that we have, at least, an option for the hot failover, those VMs that it is safe to do it with, we can get that going so that downtime of one server isn't near as critical as it once was ... ---- Marc G. Fournier Hub.Org Networking Services (http://www.hub.org) Email: scrappy@hub.org Yahoo!: yscrappy ICQ: 7615664
On Wed, 29 Sep 2004, Darcy Buskermolen wrote: >> I think you are the only 'known to work site' ... so, if I setup a round >> robin between the Panama and Canadian location for www.postgresql.org, >> would you be okay with the extra load? :) If so, I'll get to work on the >> coding changing required for the DNS ... > > I have no issues with that being setup now, I'll keep na eye on the bandwith > and let you know if it becomes problematic. 'k, will play with the DNS code tonight then ... > <script type="text/javascript"><!-- > google_ad_client = "pub-6839445463094639"; > google_ad_width = 120; > google_ad_height = 240; > google_ad_format = "120x240_as"; > google_ad_channel =""; > google_color_border = "DFF2FD"; > google_color_bg = "DFF2FD"; > google_color_link = "000000"; > google_color_url = "336699"; > google_color_text = "333333"; > //--></script> > <script type="text/javascript" > src="http://pagead2.googlesyndication.com/pagead/show_ads.js"> > </script> Sweet ... okay, will look at getting that added to archives then ... thanks for the pointer ... ---- Marc G. Fournier Hub.Org Networking Services (http://www.hub.org) Email: scrappy@hub.org Yahoo!: yscrappy ICQ: 7615664
On Wed, 29 Sep 2004, Josh Berkus wrote: > The other BSD experts I've talked to all agree that the Jail setup on > the current servers is a severe performance drain. I think we really > need to look at not using jails for the the dedicated PostgreSQL > server(s), when we have them. k, have your experts talk to my experts :) more seriously, I have two Dual-PIIIs running with ~25 jails each, and there are no performance issues that are readily apparent, either with processes or network ... even neptune, where there are currently 59 running, isn't breaking much of a sweat, to the point that our conversation here is almost real time, and that is where the mailing lists are running off of ... All the 'jail' does is a chroot to a specific directory structure, and tags the processes running on the server with the 'jail-id' so that ps inside of the jail only shows those processed appropriately tags ... unless I've been mis-informed, there is no other overhead ... once we can put dedicated servers in place, the # of jails running won't be more then 10, unless there is some major growth in requirements ... we could probably reduce the # of jails though ... pugs could be merged in with www quite easily, but I wouldn't merge cvs/pgfoundry and gborg into the same jail for what I would *hope* are obvious reasons ... If your experts can give me something more substantial to work with then opinions, though, I'd happily read it ... > FWIW, some folks at SourceForge have offered their help in configuring > this. I'm currently waiting for Adi to get back from vacation to > proceed. As there isn't much required in order to config this (as its stuff that we're already doing to load balance things like the banner ads), not quite sure what they have in mind *shrug* ---- Marc G. Fournier Hub.Org Networking Services (http://www.hub.org) Email: scrappy@hub.org Yahoo!: yscrappy ICQ: 7615664
Marc G. Fournier wrote: > On Wed, 29 Sep 2004, Mitch Pirtle wrote: > > > Hi Bruce, we met in Manhattan for your NYPHP presentation - I was the > > lone PostgreSQL cheerleader in the back ;-) > > > > You ask for proposals, no matter how radical, so here goes: > > > > o Location of servers > > > > Pending bandwidth requirements, I am willing to sponsor hosting of > > these sites, possibly on dedicated hardware. But I need some ideas of > > bandwidth usage so I can figure out which strategies make the most > > sense financially. BTW be glad your servers are not in Florida, I > > have one client that is in a significant amount of pain right now. > > The server location is not changing, but thank you for your offer ... > > > o FreeBSD > > > > I'm agnostic in this regard, but have experienced that Linux distros > > seem to be the easiest to maintain (Debian, Fedora). > > Again, operating system isn't changing, but thank you for your offer ... I assume you mean "not changing" while you are administering the servers. I don't remember any community agreement that the current setup was permanent, and my thinking now is that all avenues are open for exploration. I am not ready to dismiss this offer/idea. -- Bruce Momjian | http://candle.pha.pa.us pgman@candle.pha.pa.us | (610) 359-1001 + If your life is a hard drive, | 13 Roberts Road + Christ can be your backup. | Newtown Square, Pennsylvania 19073
Devrim GUNDUZ wrote: > I hope we'll not begin a OS/distro war! (O.K, it'd be good if Red Hat > contributes RHEL to PostgreSQL project ;) Anyone from Red Hat in here, or > in -core? ;) ) I have to say I am about as pro-*BSD as they come, but if Linux is what it will take to make things better, I am ready to try it. If it takes SCO Unixware, I am ready to try it. However, I think Marc has pointed out some less drastic changes that will fix the fsck problem. My point is that the fsck problem has gone on too long and it is time to address it head-on. Let's take the approach with the best possibility of success and implement it. > > o Web site > > > > We have been talking about a new web page layout for years at this > > point. I almost don't care if they just put a dancing bear up on the > > web site. Let's do something! > > Like this one? : http://www.jensummerfield.supanet.com/bearaniy.gif Yes, and its animated too. Everyone knows animation on a web site spices it up, as does blinking text and text that changes color. :-) > Well, web team (mostly Alexey) is working on it. We have a fabulous new > design (you know...). I think it'll be ready before 8.0.0. See the -www > archives for that (Archives? Umm maybe Google...) Yes, but how far is it from the image I saw to the completed site? I have no idea, and why has there been so little progress in the past few years? > > o Archives > > > > The archives situation is a continual problem. Again, maybe a dancing > > bear can help. :-) > > :-) A Google search box could be better sometimes... Yes, except Google has decided to stop archiving our groups since September 16. We reported this on the www list, I think. I think they had the best archive interface, and now it is gone. Take a look: http://groups.google.com/groups?hl=en&lr=&ie=UTF-8&c2coff=1&group=comp.databases.postgresql.hackers -- Bruce Momjian | http://candle.pha.pa.us pgman@candle.pha.pa.us | (610) 359-1001 + If your life is a hard drive, | 13 Roberts Road + Christ can be your backup. | Newtown Square, Pennsylvania 19073
Bruce Momjian <pgman@candle.pha.pa.us> writes:
> Yes, but how far is it from the image I saw to the completed site? I
> have no idea, and why has there been so little progress in the past few
> years?
Lack of people doing the work is why. Complaining that nothing is
getting done is just sucking time away from more productive activities
... and I think that extends to all the points made in this thread.
We all know perfectly well what the problems are --- it's a matter of
finding time, and sometimes money, to solve them. We don't need "new
ideas" unless the ideas are about how to produce those resources.
regards, tom lane
Marc G. Fournier wrote:
> On Wed, 29 Sep 2004, Bruce Momjian wrote:
>
> > It is my opinion that we have to make major changes in the way we
> > provide hosting for our servers. There are several problems:
> >
> > o Location of servers
> >
> > The location of our servers in Panama is a problem. They are too far
> > for any PostgreSQL maintainers to access. Changing hardware or
> > diagnosing problems has been too hard. I have had like 2 days of
> > downtime on my home machine in the past 12 years. We have had more than
> > 2 days of downtime in the past 6 months. My wife would not accept such
> > a reliability level.
>
> This is currently being worked on ... we are looking at various remote
> management solutions so that we don't have to deal with waiting for a
> technician to get 'on the scene' ...
OK.
> > o FreeBSD
> >
> > The use of FreeBSD jails can cause servers to take +8 hours to fsck on a
> > server crash or power failure. Again, I would never accept such
> > problems on my home server so it is hard to fathom how a project with
> > thousands of users can accept that. Either we need to find a fix, stop
> > using jails, or get another operating system, but continuing to use a
> > setup with a known problem is just asking for trouble.
>
> Actually, again, this one is being addressed ... there is a solution in
> the pipeline to fix the cause of the 8+ hour fsck, but, since it is a fix
> to fsck itself, it hasn't been put into the mainstream code yet, due to
> *obvious* testing reasons ...
Even if it is experimental, I think we should try it. But the larger
issue is why we are having OS crashes in the first place. Not all the
fsck downtime was because of power failure. What are the causes of the
other downtime, and if we don't know or can't fix it, I think we need to
look at more dramatic changes to increase reliability.
> We've also added in hot failover as an option ... I've posted to -www
> asking about putting www.postgresql.org onto it, but so far, the only
> responses back have been along the lines of 'how are you doing it?' ...
>
> The "risk" is that its not real time replication between the live and
> failover server ... on our high performance servers, the 'delay' is about
> 5 minutes ...
>
> ... now, knowing that, if you feel comfortable with me putting this onto
> the mailing lists/cvsroot as well, knowing that there is a possibility of
> something being written 'in the gap' before failover, I'll do that VM also
> ...
>
> note that altho the replication has a gap, the heartbeat process runs
> every minute ... as soon as it can't ping anymore, it fails over ...
Right, failover is nice, but we need to find the cause of why we need
the failover so often. From my perspective, if your servers can't be as
reliable as my home machine, there is something fundementally wrong,
either in hardware purchase, hosting provider, operating system,
software infrastructure, something. I don't know what the problem is,
but I know a problem when I see it.
> > o Web site
> >
> > We have been talking about a new web page layout for years at this
> > point. I almost don't care if they just put a dancing bear up on the
> > web site. Let's do something!
>
> What's wrong with the existing one? Have you designed the dancing bear
> you'd like us to put up in place of what we have now?
Looking around now. Perhaps a dancing elephant. WARNING: This will
make you ill:
http://janetskiles.com/ART/greeting/greet-ani/dancing-elephant.jpg
:-)
> > The archives situation is a continual problem. Again, maybe a dancing
> > bear can help. :-)
>
> What is wrong with it now? I'm cleaning up the code itself, but that is
> due to it being a mess right now, not due to any problems reported to me,
> removed one of the banner ads so that loading is a bit faster, and John
> has done, I think, a fantastic job on the search engine itself, including
> sending me changes for the archives themselves so that the 'time searches'
> should now work properly ...
>
> So, do you have something specific you'd like to point out to us that
> we've overlooked and haven't fixed yet?
"It not working" seems to be a continual problem. I don't know the
details and maybe it is already being fixed.
> > Basically, with no money and no one offering servers
>
> So far, I've had one person donate $10 ... in order to put a dedicated
> server onto the network, I'd need alot more of those ... that would pretty
> much eliminate your second point about the fsck's, since its only our
> *loaded* servers, that we have that problem with ... but, as I said, the
> fsck issue is being addressed as well ...
Yep, that is pretty pathetic.
--
Bruce Momjian | http://candle.pha.pa.us
pgman@candle.pha.pa.us | (610) 359-1001
+ If your life is a hard drive, | 13 Roberts Road
+ Christ can be your backup. | Newtown Square, Pennsylvania 19073
Josh Berkus wrote: > The big obstacle in moving anything is bandwidth. While any number > of companies will offer to host stuff for us, we can't match the > amount of bandwidth we're currently using in Panama ... which amounts to exactly what? -- Peter Eisentraut http://developer.postgresql.org/~petere/
I took a look at the google ad site. There is a lot more to it than the part where you put the java script onto the page. You have to create an account and be approved, etc. etc. etc. --elein On Wed, Sep 29, 2004 at 04:13:05PM -0300, Marc G. Fournier wrote: > On Wed, 29 Sep 2004, Darcy Buskermolen wrote: > > >>I think you are the only 'known to work site' ... so, if I setup a round > >>robin between the Panama and Canadian location for www.postgresql.org, > >>would you be okay with the extra load? :) If so, I'll get to work on the > >>coding changing required for the DNS ... > > > >I have no issues with that being setup now, I'll keep na eye on the > >bandwith > >and let you know if it becomes problematic. > > 'k, will play with the DNS code tonight then ... > > ><script type="text/javascript"><!-- > >google_ad_client = "pub-6839445463094639"; > >google_ad_width = 120; > >google_ad_height = 240; > >google_ad_format = "120x240_as"; > >google_ad_channel =""; > >google_color_border = "DFF2FD"; > >google_color_bg = "DFF2FD"; > >google_color_link = "000000"; > >google_color_url = "336699"; > >google_color_text = "333333"; > >//--></script> > ><script type="text/javascript" > > src="http://pagead2.googlesyndication.com/pagead/show_ads.js"> > ></script> > > Sweet ... okay, will look at getting that added to archives then ... > thanks for the pointer ... > > ---- > Marc G. Fournier Hub.Org Networking Services (http://www.hub.org) > Email: scrappy@hub.org Yahoo!: yscrappy ICQ: 7615664 > > ---------------------------(end of broadcast)--------------------------- > TIP 3: if posting/reading through Usenet, please send an appropriate > subscribe-nomail command to majordomo@postgresql.org so that your > message can get through to the mailing list cleanly
Tom Lane wrote: > We all know perfectly well what the problems are --- > it's a matter of finding time, and sometimes money, to solve them. > We don't need "new ideas" unless the ideas are about how to produce > those resources. Throwing more resources at an entity that does not scale won't achieve anything. In order to improve, we need to listen to competing concepts and pick the best one, not keep giving the current one more time over and over. -- Peter Eisentraut http://developer.postgresql.org/~petere/
On Wed, 29 Sep 2004, elein wrote: > I took a look at the google ad site. There is a lot more to it > than the part where you put the java script onto the page. > You have to create an account and be approved, etc. etc. etc. yup, already put through the application, Darcy says it takes about 48 hours or so to be approved ... > > --elein > > > On Wed, Sep 29, 2004 at 04:13:05PM -0300, Marc G. Fournier wrote: >> On Wed, 29 Sep 2004, Darcy Buskermolen wrote: >> >>>> I think you are the only 'known to work site' ... so, if I setup a round >>>> robin between the Panama and Canadian location for www.postgresql.org, >>>> would you be okay with the extra load? :) If so, I'll get to work on the >>>> coding changing required for the DNS ... >>> >>> I have no issues with that being setup now, I'll keep na eye on the >>> bandwith >>> and let you know if it becomes problematic. >> >> 'k, will play with the DNS code tonight then ... >> >>> <script type="text/javascript"><!-- >>> google_ad_client = "pub-6839445463094639"; >>> google_ad_width = 120; >>> google_ad_height = 240; >>> google_ad_format = "120x240_as"; >>> google_ad_channel =""; >>> google_color_border = "DFF2FD"; >>> google_color_bg = "DFF2FD"; >>> google_color_link = "000000"; >>> google_color_url = "336699"; >>> google_color_text = "333333"; >>> //--></script> >>> <script type="text/javascript" >>> src="http://pagead2.googlesyndication.com/pagead/show_ads.js"> >>> </script> >> >> Sweet ... okay, will look at getting that added to archives then ... >> thanks for the pointer ... >> >> ---- >> Marc G. Fournier Hub.Org Networking Services (http://www.hub.org) >> Email: scrappy@hub.org Yahoo!: yscrappy ICQ: 7615664 >> >> ---------------------------(end of broadcast)--------------------------- >> TIP 3: if posting/reading through Usenet, please send an appropriate >> subscribe-nomail command to majordomo@postgresql.org so that your >> message can get through to the mailing list cleanly > ---- Marc G. Fournier Hub.Org Networking Services (http://www.hub.org) Email: scrappy@hub.org Yahoo!: yscrappy ICQ: 7615664
Marc, > yup, already put through the application, Darcy says it takes about 48 > hours or so to be approved ... Um, can we vote on this? It's my long-term goal to get rid of all ads on postgresql.org, so I'd like to hunt for other funding sources before doing this. -- --Josh Josh Berkus Aglio Database Solutions San Francisco
Hi, Bruce Momjian wrote: >>Like this one? : http://www.jensummerfield.supanet.com/bearaniy.gif > > Yes, and its animated too. Everyone knows animation on a web site > spices it up, as does blinking text and text that changes color. :-) You forgot background music in MIDI. >>Well, web team (mostly Alexey) is working on it. We have a fabulous new >>design (you know...). I think it'll be ready before 8.0.0. See the -www >>archives for that (Archives? Umm maybe Google...) > > Yes, but how far is it from the image I saw to the completed site? I > have no idea, and why has there been so little progress in the past few > years? Er, the "beta" site is already working: http://alexey.beta.postgresql.org/ I am unfortunately a bit busy ATM but expect to continue work on it the next week. There is very little work left, in fact. As for the "past few years" --- I've no idea. I volunteered only this year and had to overcome some resistance when I said that the current site is inadequate.
-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 Hi, On Wed, 29 Sep 2004, Bruce Momjian wrote: >> :-) A Google search box could be better sometimes... > > Yes, except Google has decided to stop archiving our groups since > September 16. We reported this on the www list, I think. I think they > had the best archive interface, and now it is gone. Take a look: > > http://groups.google.com/groups?hl=en&lr=&ie=UTF-8&c2coff=1&group=comp.databases.postgresql.hackers Dave has contacted with Google, and I'm sure that they'll solve this problem. See: http://archives.postgresql.org/pgsql-www/2004-09/msg00191.php Regards, - -- Devrim GUNDUZ devrim~gunduz.org devrim.gunduz~linux.org.tr http://www.tdmsoft.com http://www.gunduz.org -----BEGIN PGP SIGNATURE----- Version: GnuPG v1.2.1 (GNU/Linux) iD8DBQFBWzD8tl86P3SPfQ4RAiiYAJ9JOO2n0E55slW8e8STLCGfej4P8ACg19zk 6smnsBQgIUxTWRDlj9TimHE= =JCZW -----END PGP SIGNATURE-----
Peter Eisentraut wrote: > Josh Berkus wrote: > >>The big obstacle in moving anything is bandwidth. While any number >>of companies will offer to host stuff for us, we can't match the >>amount of bandwidth we're currently using in Panama > > ... which amounts to exactly what? If it's helpful, I can ask my boss here if we're willing to host the PG project. Lots of bandwidth and grunty servers. Any rough ballpark figures on the amount of bandwidth we'd have to figure in? i.e. xxx per month? Regards and best wishes, Justin Clift
On Wed, 29 Sep 2004, Alvaro Herrera wrote: > about pg_query not getting a good connection or similar problems. > Isn't normal advice to check return codes before even trying to use the > connections? Why isn't this done in the main postgresql.org code, where > anyone can see it, is beyond me. Of course the solution to the > underlying problem is to restart the Postgres server, but why should we > inform the user that Postgres' own database server is down, in the worst > possible way? Just curious here, but when/where? We haven't had a database issue in quite awhile that *I'm* aware of ... in fact, the whole web site is static, generated periodically from .php files, so that the mirrors can pick things up properly, so there should never be a 'cannot connect to database' issue, since there are no connections to the database being made ... ---- Marc G. Fournier Hub.Org Networking Services (http://www.hub.org) Email: scrappy@hub.org Yahoo!: yscrappy ICQ: 7615664
Fixed ... On Wed, 29 Sep 2004, Alvaro Herrera wrote: > On Thu, Sep 30, 2004 at 01:02:33AM +0300, Devrim GUNDUZ wrote: > >> Dave has contacted with Google, and I'm sure that they'll solve this >> problem. See: >> >> http://archives.postgresql.org/pgsql-www/2004-09/msg00191.php > > Now that people are talking about www and archives, please take a look > at a problem I've pointed a couple of times in the past and had no > answer: the pgsql-es-ayuda list is not on the left column. It's also > not listed on the " Search in: " combo box at the top either. > > http://archives.postgresql.org/pgsql-www/2004-09/msg00191.php > > Why is this? Any chance to correct it? > > Note that, of the regional lists, this is the one with the most traffic, > and the oldest (don't be fooled by the length of the month list ... the > last item has articles going back to 1997, it's just badly built) > > -- > Alvaro Herrera (<alvherre[a]dcc.uchile.cl>) > "Cuando no hay humildad las personas se degradan" (A. Christie) > > > ---------------------------(end of broadcast)--------------------------- > TIP 3: if posting/reading through Usenet, please send an appropriate > subscribe-nomail command to majordomo@postgresql.org so that your > message can get through to the mailing list cleanly > ---- Marc G. Fournier Hub.Org Networking Services (http://www.hub.org) Email: scrappy@hub.org Yahoo!: yscrappy ICQ: 7615664
On Wed, 29 Sep 2004, Alvaro Herrera wrote: > On Wed, Sep 29, 2004 at 08:35:51PM -0300, Marc G. Fournier wrote: >> >> Fixed ... > > Thanks ... there is a typo though: on the upper menu, the combo box says > "German" where it should say "Spanish". Oops, sorry, was even looking for that and missed it :( Fixed now ... > >> On Wed, 29 Sep 2004, Alvaro Herrera wrote: >> >>> On Thu, Sep 30, 2004 at 01:02:33AM +0300, Devrim GUNDUZ wrote: >>> >>>> Dave has contacted with Google, and I'm sure that they'll solve this >>>> problem. See: >>>> >>>> http://archives.postgresql.org/pgsql-www/2004-09/msg00191.php >>> >>> Now that people are talking about www and archives, please take a look >>> at a problem I've pointed a couple of times in the past and had no >>> answer: the pgsql-es-ayuda list is not on the left column. It's also >>> not listed on the " Search in: " combo box at the top either. > > -- > Alvaro Herrera (<alvherre[a]dcc.uchile.cl>) > "God is real, unless declared as int" > > ---- Marc G. Fournier Hub.Org Networking Services (http://www.hub.org) Email: scrappy@hub.org Yahoo!: yscrappy ICQ: 7615664
> -----Original Message----- > From: pgsql-www-owner@postgresql.org > [mailto:pgsql-www-owner@postgresql.org] On Behalf Of Devrim GUNDUZ > Sent: 29 September 2004 18:26 > To: Bruce Momjian > Cc: PostgreSQL WWW Mailing List; PostgreSQL advocacy > Subject: Re: [pgsql-www] Server unreliability > > :-) A Google search box could be better sometimes... Not now they've stopped archiving the largest chunk of our sites - namely the lists. /D
On September 29, 2004 10:13 am, you wrote: > It is my opinion that we have to make major changes in the way we > provide hosting for our servers. There are several problems: > > o Location of servers > > The location of our servers in Panama is a problem. They are too far > for any PostgreSQL maintainers to access. Changing hardware or > diagnosing problems has been too hard. I have had like 2 days of > downtime on my home machine in the past 12 years. We have had more than > 2 days of downtime in the past 6 months. My wife would not accept such > a reliability level. Yes location is a bit of a concern, but I think it's muchless of a concern than the rest of the issues. I think with hardware/jail issues sorted out this one will for the most part take careof it's self, or at least be far less apparent. > > o FreeBSD > > The use of FreeBSD jails can cause servers to take +8 hours to fsck on a > server crash or power failure. Again, I would never accept such > problems on my home server so it is hard to fathom how a project with > thousands of users can accept that. Either we need to find a fix, stop > using jails, or get another operating system, but continuing to use a > setup with a known problem is just asking for trouble. The over use of jails I think is complicating things, what are the odds that someone could find it in their heart to raid the old hardware pile in their serverback room, and come up with a PIII 800MHz + system with a gig of ram that they could donate to the cause, and we move some of the more critical services (www/ftp/anoncvs) to that box without jails. Yahoo, Pair, Fugtisu, Affilias , Somebody, Anybody??? > > o Web site > > We have been talking about a new web page layout for years at this > point. I almost don't care if they just put a dancing bear up on the > web site. Let's do something! One thing I know we can do here to help at least the availability of the website itself, is to use a round-robin DNS based setup. I know Marc adjusted DNS records to point the main website at www3.ca a while back during a time that www was stuck in fsck hell. As for the website redesign, I agree fully, it's been talked about for at least 3 full releases now and from my understanding it's no closer now than it was before. Can anybody comment on the hold up here? > > o Archives > > The archives situation is a continual problem. Again, maybe a dancing > bear can help. :-) What about combining the efforts od pgsql.ru and command prompt to do this, or maybe even monarc or google groups ? > > Basically, with no money and no one offering servers, I don't see a good > solution to any of these problems, but I think we need to recognize > these are problems and that we will continue to suffer until they are > addressed. Ok i know how much we all hate banner adds and the like, but what about using Google ads especially on the archives to help generate income to fund these requirements? > > Are there any proposals, no matter how radical, to correct these? -- Darcy Buskermolen Wavefire Technologies Corp. ph: 250.717.0200 fx: 250.763.1759 http://www.wavefire.com
Bruce Momjian wrote: >It is my opinion that we have to make major changes in the way we >provide hosting for our servers. There are several problems: > >o Location of servers > >The location of our servers in Panama is a problem. They are too far >for any PostgreSQL maintainers to access. Changing hardware or >diagnosing problems has been too hard. I have had like 2 days of >downtime on my home machine in the past 12 years. We have had more than >2 days of downtime in the past 6 months. My wife would not accept such >a reliability level. > > Could somebody explain to me how panama came to be chosen. Those reasons may have changed since then. >o FreeBSD > >The use of FreeBSD jails can cause servers to take +8 hours to fsck on a >server crash or power failure. Again, I would never accept such >problems on my home server so it is hard to fathom how a project with >thousands of users can accept that. Either we need to find a fix, stop >using jails, or get another operating system, but continuing to use a >setup with a known problem is just asking for trouble. > > I'm very comfortable using Debian and would be willing to remote admin using that. As for sticking with FreeBSD its the better choice if there's somebody up to speed with security. The best resource person I know is OReilly's FreeBSD columnist Dru, dlavigne6.sympatico.ca. She might be convinced to "advise" what to do. >o Archives > >The archives situation is a continual problem. Again, maybe a dancing >bear can help. :-) > > How does the archives work i.e. what's running it? I hate to state the obvious but can we have an itemized list as to what's wrong with them. I've got my own ideas but I'd like to be on the same page as everybody else. >Are there any proposals, no matter how radical, to correct these? > > I've got thoughts as to server hosting but I don't even want to discuss it until the guts of the site has been updated. Robert
> -----Original Message----- > From: pgsql-www-owner@postgresql.org > [mailto:pgsql-www-owner@postgresql.org] On Behalf Of Marc G. Fournier > Sent: 29 September 2004 18:53 > To: Bruce Momjian > Cc: PostgreSQL www; PostgreSQL advocacy > Subject: Re: [pgsql-www] Server unreliability > > We've also added in hot failover as an option ... I've posted > to -www asking about putting www.postgresql.org onto it, but > so far, the only responses back have been along the lines of > 'how are you doing it?' ... Err, no. I checked what's on WWW and told you it wasn't a problem, go ahead. Regards, Dave.
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1
Hi,
On Wed, 29 Sep 2004, Mark Harrison wrote:
> 1. Move primary file downloads to sourceforge. That should reduce the
> traffic load a lot.
>
> 2. Move the mailing lists to a mailing list server (e.g. yahoo).
Are you kidding?
> 4. For the website, concentrate on simple static pages that can
> be served efficiently.
That's what we are doing now...
Regards,
- --
Devrim GUNDUZ
devrim~gunduz.org devrim.gunduz~linux.org.tr
http://www.tdmsoft.com
http://www.gunduz.org
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.1 (GNU/Linux)
iD4DBQFBW7Nytl86P3SPfQ4RAtdFAJdpoosF8oW3c4nCyK/Gc2AcibFBAKCIkFGK
9d3qctBlMKOfH8vXu3xhiQ==
=18UW
-----END PGP SIGNATURE-----
Hi Bruce, we met in Manhattan for your NYPHP presentation - I was the lone PostgreSQL cheerleader in the back ;-) You ask for proposals, no matter how radical, so here goes: o Location of servers Pending bandwidth requirements, I am willing to sponsor hosting of these sites, possibly on dedicated hardware. But I need some ideas of bandwidth usage so I can figure out which strategies make the most sense financially. BTW be glad your servers are not in Florida, I have one client that is in a significant amount of pain right now. o FreeBSD I'm agnostic in this regard, but have experienced that Linux distros seem to be the easiest to maintain (Debian, Fedora). o Web site I am a core developer on the Mambo project (www.mamboserver.com), and joined the core team primarily to implement database abstraction (currently tied to MySQL gasp!). Mambo has grown quite famous for being easy to use, has lightweight requirements, and has a huge 3rd party community that provides all kinds of extra goodies (forums, i18n support, commerce, you name it). I would happily volunteer the efforts to migrate the existing site to Mambo to make it easier for the core gang to keep the site current. I would also consider having PostgreSQL.org running Mambo being excellent incentive to prioritize the implementation of ADOdb ;-) o Archives Depending on the format, Mambo natively supports archiving of content. This may be a big win for PostgreSQL; and I would just need to get more information on the scope (and current format) of the archived content. o About me I have been very active as a PostgreSQL booster (not as a developer), so none of you have ever heard of me - but I have been very vocal in both Switzerland (where I lived until last year) and in Manhattan. Writing articles for International PHP Magazine (www.php-mag.net) also gives me a chance for some subliminal advertising. There you have it, consider it as a preliminary offer - it is the best I can do, being that all of you have given me such a powerful database for free. Perhaps being a web developer means that the best way for me to contribute would be on the website? Regards, Mitch Pirtle Mambo core eveloper (www.mamboserver.com) International PHP Magazine author (www.php-mag.net)
On September 29, 2004 11:09 am, Marc G. Fournier wrote: > On Wed, 29 Sep 2004, Darcy Buskermolen wrote: ----8<----[snip] > jails are not the problem ... the problem is the file system we are using > to share applications ... but, yes, if we could put a dedicated box in > place, we could very easily eliminate the use of the file system ... we > have never had a jail related crash on those servers, they have always > come down to the file system itself ... Poor use of terminlogy on my part, but the net effect is the same, the FS is being used to facilitate the jail, take away the jail, take away the need of the stacked FS. > To do this, and I'd love to, we'd need to have *at least* all of the > www*.us.postgresql.org mirrors setup a ServerAlias for www.postgresql.org > ... not sure what would be involved in getting that done though, but the > DNS side wouldn't take much ... *nod* In the interim there is nothing stoping getting the process rolling with those sites that are known to work, as well as contacting the mirror admins. of the unknown sites.. > > > What about combining the efforts od pgsql.ru and command prompt to do > > this, or maybe even monarc or google groups ? > > Until someone can *list* the problems, how can a solution be suggested? > John and I know of no issues at this time ... the last one was the 'time > period' searches, which we believe we've just fixed ... I'm not sure what the issues have been other than availability. > > > Ok i know how much we all hate banner adds and the like, but what about > > using Google ads especially on the archives to help generate income to > > fund these requirements? > > Huh? Google ads? I know what you mean, just how do we make use of them? Specificly the program out lines at: https://www.google.com/adsense/ I know of a few reasonably high traffic sites that are generating about 5k/mth income from this program. > > ---- > Marc G. Fournier Hub.Org Networking Services (http://www.hub.org) > Email: scrappy@hub.org Yahoo!: yscrappy ICQ: 7615664 > > ---------------------------(end of broadcast)--------------------------- > TIP 2: you can get off all lists at once with the unregister command > (send "unregister YourEmailAddressHere" to majordomo@postgresql.org) -- Darcy Buskermolen Wavefire Technologies Corp. ph: 250.717.0200 fx: 250.763.1759 http://www.wavefire.com
On September 29, 2004 11:49 am, Marc G. Fournier wrote: > On Wed, 29 Sep 2004, Darcy Buskermolen wrote: > > On September 29, 2004 11:09 am, Marc G. Fournier wrote: > >> On Wed, 29 Sep 2004, Darcy Buskermolen wrote: ----8<----[snip] > > *nod* In the interim there is nothing stoping getting the process > > rolling with those sites that are known to work, as well as contacting > > the mirror admins. of the unknown sites.. > > I think you are the only 'known to work site' ... so, if I setup a round > robin between the Panama and Canadian location for www.postgresql.org, > would you be okay with the extra load? :) If so, I'll get to work on the > coding changing required for the DNS ... I have no issues with that being setup now, I'll keep na eye on the bandwith and let you know if it becomes problematic. > > > Specificly the program out lines at: https://www.google.com/adsense/ I > > know of a few reasonably high traffic sites that are generating about > > 5k/mth income from this program. > > 'k, will have to read through, but first scan, would the replace our > current search system, or would the 'Google search box' be on the right of > the screen, or ... ? Without me reading it all the way through, do you > already know? Not the search system, but the addsence. it's just a banner serving technology where you embed piece of javascript to imbed in the site and away we go.. For example here is a sample I just fired up based on what I use in another site.... http://www.dbitech.ca/ and the JS to pull it of is... <script type="text/javascript"><!-- google_ad_client = "pub-6839445463094639"; google_ad_width = 120; google_ad_height = 240; google_ad_format = "120x240_as"; google_ad_channel =""; google_color_border = "DFF2FD"; google_color_bg = "DFF2FD"; google_color_link = "000000"; google_color_url = "336699"; google_color_text = "333333"; //--></script> <script type="text/javascript" src="http://pagead2.googlesyndication.com/pagead/show_ads.js"> </script> > > ---- > Marc G. Fournier Hub.Org Networking Services (http://www.hub.org) > Email: scrappy@hub.org Yahoo!: yscrappy ICQ: 7615664 -- Darcy Buskermolen Wavefire Technologies Corp. ph: 250.717.0200 fx: 250.763.1759 http://www.wavefire.com
On Wed, Sep 29, 2004 at 04:07:20PM -0300, Marc G. Fournier wrote: > as to what advantages the jail's provide ... mainly, security ... we can > easily provide root access to pugs.postgresql.org, as an example, so that > Fred (please tell me I remembered right?) can do whatever work he wants, > without giving him full root on the developer jail ... same with all of > the other jails ... Chris has full root on gborg, so that he doesn't have > to ask someone else to do something when he notices a problem (ie. restart > mailman) ... This business about anyone having to restart anything is strange. I don't run mailman lists personally anymore, but back when I used to, there was no need to restart anything. This sounds like there are other reliability problems that should be attacked. Also I understand that it's a problem that fsck takes 8 hours to run. However, why on earth is fsck running in the first place? Do the servers go down unexpectedly on regular basis? This also screams of a more fundamental problem that should be solved. Another thing: I've seen several times already PHP warnings show up about pg_query not getting a good connection or similar problems. Isn't normal advice to check return codes before even trying to use the connections? Why isn't this done in the main postgresql.org code, where anyone can see it, is beyond me. Of course the solution to the underlying problem is to restart the Postgres server, but why should we inform the user that Postgres' own database server is down, in the worst possible way? Just random thoughts ... -- Alvaro Herrera (<alvherre[a]dcc.uchile.cl>) "Hay dos momentos en la vida de un hombre en los que no debería especular: cuando puede permitírselo y cuando no puede" (Mark Twain)
> -----Original Message----- > From: pgsql-www-owner@postgresql.org > [mailto:pgsql-www-owner@postgresql.org] On Behalf Of Bruce Momjian > Sent: 29 September 2004 21:18 > To: Devrim GUNDUZ > Cc: PostgreSQL WWW Mailing List; PostgreSQL advocacy > Subject: Re: [pgsql-www] Server unreliability > > Yes, but how far is it from the image I saw to the completed > site? http://wwwdevel.postgresql.org/ > I have no idea, and why has there been so little > progress in the past few years? The current site is only a little over 18 months old iirc. I'd hardly call that little progress in years... Regards, Dave
> -----Original Message----- > From: pgsql-www-owner@postgresql.org > [mailto:pgsql-www-owner@postgresql.org] On Behalf Of Devrim GUNDUZ > Sent: 29 September 2004 23:03 > To: Bruce Momjian > Cc: PostgreSQL WWW Mailing List; PostgreSQL advocacy > Subject: Re: [pgsql-www] Server unreliability > > -----BEGIN PGP SIGNED MESSAGE----- > Hash: SHA1 > > > Hi, > > On Wed, 29 Sep 2004, Bruce Momjian wrote: > > >> :-) A Google search box could be better sometimes... > > > > Yes, except Google has decided to stop archiving our groups since > > September 16. We reported this on the www list, I think. I think > > they had the best archive interface, and now it is gone. > Take a look: > > > > > > > http://groups.google.com/groups?hl=en&lr=&ie=UTF-8&c2coff=1&group=comp > > .databases.postgresql.hackers > > Dave has contacted with Google, and I'm sure that they'll > solve this problem. See: It's not looking good. I think we're going to need a mass of complaints form the users. Regards, Dave
On Thu, Sep 30, 2004 at 01:02:33AM +0300, Devrim GUNDUZ wrote: > Dave has contacted with Google, and I'm sure that they'll solve this > problem. See: > > http://archives.postgresql.org/pgsql-www/2004-09/msg00191.php Now that people are talking about www and archives, please take a look at a problem I've pointed a couple of times in the past and had no answer: the pgsql-es-ayuda list is not on the left column. It's also not listed on the " Search in: " combo box at the top either. http://archives.postgresql.org/pgsql-www/2004-09/msg00191.php Why is this? Any chance to correct it? Note that, of the regional lists, this is the one with the most traffic, and the oldest (don't be fooled by the length of the month list ... the last item has articles going back to 1997, it's just badly built) -- Alvaro Herrera (<alvherre[a]dcc.uchile.cl>) "Cuando no hay humildad las personas se degradan" (A. Christie)
On Wed, Sep 29, 2004 at 08:35:51PM -0300, Marc G. Fournier wrote: > > Fixed ... Thanks ... there is a typo though: on the upper menu, the combo box says "German" where it should say "Spanish". > On Wed, 29 Sep 2004, Alvaro Herrera wrote: > > >On Thu, Sep 30, 2004 at 01:02:33AM +0300, Devrim GUNDUZ wrote: > > > >>Dave has contacted with Google, and I'm sure that they'll solve this > >>problem. See: > >> > >>http://archives.postgresql.org/pgsql-www/2004-09/msg00191.php > > > >Now that people are talking about www and archives, please take a look > >at a problem I've pointed a couple of times in the past and had no > >answer: the pgsql-es-ayuda list is not on the left column. It's also > >not listed on the " Search in: " combo box at the top either. -- Alvaro Herrera (<alvherre[a]dcc.uchile.cl>) "God is real, unless declared as int"
Bruce Momjian wrote:
> It is my opinion that we have to make major changes in the way we
> provide hosting for our servers. There are several problems:
>
> o Location of servers
>
> The location of our servers in Panama is a problem. They are too far
> for any PostgreSQL maintainers to access. Changing hardware or
> diagnosing problems has been too hard. I have had like 2 days of
> downtime on my home machine in the past 12 years. We have had more than
> 2 days of downtime in the past 6 months. My wife would not accept such
> a reliability level.
>
> o FreeBSD
>
> The use of FreeBSD jails can cause servers to take +8 hours to fsck on a
> server crash or power failure. Again, I would never accept such
> problems on my home server so it is hard to fathom how a project with
> thousands of users can accept that. Either we need to find a fix, stop
> using jails, or get another operating system, but continuing to use a
> setup with a known problem is just asking for trouble.
>
> o Web site
>
> We have been talking about a new web page layout for years at this
> point. I almost don't care if they just put a dancing bear up on the
> web site. Let's do something!
>
> o Archives
>
> The archives situation is a continual problem. Again, maybe a dancing
> bear can help. :-)
>
> Basically, with no money and no one offering servers, I don't see a good
> solution to any of these problems, but I think we need to recognize
> these are problems and that we will continue to suffer until they are
> addressed.
>
> Are there any proposals, no matter how radical, to correct these?
Here are a few radical ideas... The basic idea is to leverage as many
outside resources as possible.
1. Move primary file downloads to sourceforge. That should reduce the
traffic load a lot.
2. Move the mailing lists to a mailing list server (e.g. yahoo).
3. Base the new website on a blog-style package that makes it
easy to update for multiple people (disclaimer, I don't know
what the current site is running, so maybe that's already
been done).
4. For the website, concentrate on simple static pages that can
be served efficiently.
5. I like the google ads idea, it will possibly generate a bit
of money and connect people with organizations interested
in postgresql.
All just IMHO, and with immense gratitude for all the work
that's already been done!
Mark
--
Mark Harrison
Pixar Animation Studios
Hello, >o Archives > >The archives situation is a continual problem. Again, maybe a dancing >bear can help. :-) > > This is being resolved by Command Prompt. Marc and I are in the process of moving archives as we speak. In fact DNS should be propagating by morning. >Basically, with no money and no one offering servers, I don't see a good >solution to any of these problems, but I think we need to recognize >these are problems and that we will continue to suffer until they are >addressed. > > > Command Prompt has offered servers and bandwidth on Multiple occassions. We have top notch redundant bandwidth, natural gas fired generators, and solid uptime. >Are there any proposals, no matter how radical, to correct these? > > > Please see above. Sincerely, Joshua D. Drake -- Command Prompt, Inc., home of Mammoth PostgreSQL - S/ODBC and S/JDBC Postgresql support, programming shared hosting and dedicated hosting. +1-503-667-4564 - jd@commandprompt.com - http://www.commandprompt.com PostgreSQL Replicator -- production quality replication for PostgreSQL
On Wed, Sep 29, 2004 at 09:34:21PM -0300, Marc G. Fournier wrote: > On Wed, 29 Sep 2004, Alvaro Herrera wrote: > > >On Wed, Sep 29, 2004 at 08:35:51PM -0300, Marc G. Fournier wrote: > >> > >>Fixed ... > > > >Thanks ... there is a typo though: on the upper menu, the combo box says > >"German" where it should say "Spanish". > > Oops, sorry, was even looking for that and missed it :( > > Fixed now ... Thanks. -- Alvaro Herrera (<alvherre[a]dcc.uchile.cl>) "Escucha y olvidarás; ve y recordarás; haz y entenderás" (Confucio)
> -----Original Message----- > From: pgsql-www-owner@postgresql.org > [mailto:pgsql-www-owner@postgresql.org] On Behalf Of Alexey Borzov > Sent: 29 September 2004 22:38 > To: Bruce Momjian > Cc: PostgreSQL WWW Mailing List > Subject: Re: [pgsql-www] Server unreliability > > As for the "past few years" --- I've no idea. I volunteered > only this year and had to overcome some resistance when I > said that the current site is inadequate. Err, yes. Well that wasn't so much resistance to it being changed as I'm sure you remember. Thankfully that's all behind us now though... Regards Dave
On Wed, Sep 29, 2004 at 08:24:31PM -0300, Marc G. Fournier wrote: > On Wed, 29 Sep 2004, Alvaro Herrera wrote: > > >about pg_query not getting a good connection or similar problems. > >Isn't normal advice to check return codes before even trying to use the > >connections? Why isn't this done in the main postgresql.org code, where > >anyone can see it, is beyond me. Of course the solution to the > >underlying problem is to restart the Postgres server, but why should we > >inform the user that Postgres' own database server is down, in the worst > >possible way? > > Just curious here, but when/where? I don't remember an exact date or place, just that I've seen that happen more than three times and I've always thought it was bad programming practice. It has been very annoying, mainly because it means casual visitors could see those and possible think Postgres reliability sucked. Perhaps it has been fixed since then. -- Alvaro Herrera (<alvherre[a]dcc.uchile.cl>) "The important things in the world are problems with society that we don't understand at all. The machines will become more complicated but they won't be more complicated than the societies that run them." (Freeman Dyson)
On Wed, Sep 29, 2004 at 06:20:37PM -0400, Alvaro Herrera wrote: > Now that people are talking about www and archives, please take a look > at a problem I've pointed a couple of times in the past and had no > answer: the pgsql-es-ayuda list is not on the left column. It's also > not listed on the " Search in: " combo box at the top either. Oh, another one: Here http://search.postgresql.org/archives.search?cs=utf-8&fm=on&st=20&dt=back&q=llave+foranea&ul=http%3A%2F%2Farchives.postgresql.org%2Fpgsql-es-ayuda%2F%25&dp=0&o=0&ps=10&s=date the pgsql-es-ayuda is not in the upper combo box. This is the page that's presented as result from a search (it was originally restricted to pgsql-es-ayuda). It seems really weird to have two lists of lists, one at the top when you are looking at plain archives, and other when you are looking at search results. Most certainly they should be identical! Why aren't they one and the same? Thanks, -- Alvaro Herrera (<alvherre[a]dcc.uchile.cl>) www.google.com: interfaz de línea de comando para la web.
On Thu, 30 Sep 2004, Alvaro Herrera wrote: > On Wed, Sep 29, 2004 at 06:20:37PM -0400, Alvaro Herrera wrote: > >> Now that people are talking about www and archives, please take a look >> at a problem I've pointed a couple of times in the past and had no >> answer: the pgsql-es-ayuda list is not on the left column. It's also >> not listed on the " Search in: " combo box at the top either. > > Oh, another one: Here > > http://search.postgresql.org/archives.search?cs=utf-8&fm=on&st=20&dt=back&q=llave+foranea&ul=http%3A%2F%2Farchives.postgresql.org%2Fpgsql-es-ayuda%2F%25&dp=0&o=0&ps=10&s=date > > the pgsql-es-ayuda is not in the upper combo box. This is the page > that's presented as result from a search (it was originally restricted > to pgsql-es-ayuda). > > It seems really weird to have two lists of lists, one at the top when > you are looking at plain archives, and other when you are looking at > search results. Most certainly they should be identical! Why aren't > they one and the same? search is on a different server then archives ... we try and keep the search box in sync, but John most likely didn't see your posting about it yesterday, and I haven't had a chance to update him ... ---- Marc G. Fournier Hub.Org Networking Services (http://www.hub.org) Email: scrappy@hub.org Yahoo!: yscrappy ICQ: 7615664
On Thu, 30 Sep 2004, Alvaro Herrera wrote: > I don't remember an exact date or place, just that I've seen that happen > more than three times and I've always thought it was bad programming > practice. It has been very annoying, mainly because it means casual > visitors could see those and possible think Postgres reliability sucked. > Perhaps it has been fixed since then. If you see it, please report which URL you see it on ... there has been alot of code cleanups happening, especially with Alexey's work on the backend code, but its possible that something has been missed ... you are right that it doesn't look good when visitors see it, but if nobody reports it, its not goign to get fixed either ... ---- Marc G. Fournier Hub.Org Networking Services (http://www.hub.org) Email: scrappy@hub.org Yahoo!: yscrappy ICQ: 7615664
On Wed, 29 Sep 2004, Marc G. Fournier wrote: > On Wed, 29 Sep 2004, Alvaro Herrera wrote: > > > about pg_query not getting a good connection or similar problems. > > Isn't normal advice to check return codes before even trying to use the > > connections? Why isn't this done in the main postgresql.org code, where > > anyone can see it, is beyond me. Of course the solution to the > > underlying problem is to restart the Postgres server, but why should we > > inform the user that Postgres' own database server is down, in the worst > > possible way? > > Just curious here, but when/where? We haven't had a database issue in > quite awhile that *I'm* aware of ... in fact, the whole web site is > static, generated periodically from .php files, so that the mirrors can > pick things up properly, so there should never be a 'cannot connect to > database' issue, since there are no connections to the database being made I know this isn't on the main site but: http://www.alcove.com.au/~swm/gborg.png Gavin
Chris? Some extra error checking needed, maybe? On Sat, 2 Oct 2004, Gavin Sherry wrote: > On Wed, 29 Sep 2004, Marc G. Fournier wrote: > >> On Wed, 29 Sep 2004, Alvaro Herrera wrote: >> >>> about pg_query not getting a good connection or similar problems. >>> Isn't normal advice to check return codes before even trying to use the >>> connections? Why isn't this done in the main postgresql.org code, where >>> anyone can see it, is beyond me. Of course the solution to the >>> underlying problem is to restart the Postgres server, but why should we >>> inform the user that Postgres' own database server is down, in the worst >>> possible way? >> >> Just curious here, but when/where? We haven't had a database issue in >> quite awhile that *I'm* aware of ... in fact, the whole web site is >> static, generated periodically from .php files, so that the mirrors can >> pick things up properly, so there should never be a 'cannot connect to >> database' issue, since there are no connections to the database being made > > I know this isn't on the main site but: > > http://www.alcove.com.au/~swm/gborg.png > > Gavin > ---- Marc G. Fournier Hub.Org Networking Services (http://www.hub.org) Email: scrappy@hub.org Yahoo!: yscrappy ICQ: 7615664