Обсуждение: InitDB Failure on install
> FATAL: could not create shared memory segment: No error > DETAIL: Failed system call was shmget(key=1, size=1196032, 03600). > child process was terminated by signal 1 > initdb: failed > initdb: removing cont I have had the same problem if i tried to install postgresql within an terminal-session on your windows 2000 terminal-server. Today (after a reboot) i tried it again from the console terminal, and guess what, it worked !! Now i cant recheck, but maybe this failure happens only if you are connected through an terminal-session with your server!?!? Ciao, Mario
Yep, I detect this too. The problem happens in terminal services session. Regards Rodrigo -----Mensagem original----- De: pgsql-hackers-win32-owner@postgresql.org [mailto:pgsql-hackers-win32-owner@postgresql.org]Em nome de Mario Ivankovits Enviada em: quinta-feira, 19 de agosto de 2004 14:51 Para: pgsql-hackers-win32@postgresql.org Assunto: [pgsql-hackers-win32] InitDB Failure on install > FATAL: could not create shared memory segment: No error > DETAIL: Failed system call was shmget(key=1, size=1196032, 03600). > child process was terminated by signal 1 > initdb: failed > initdb: removing cont I have had the same problem if i tried to install postgresql within an terminal-session on your windows 2000 terminal-server. Today (after a reboot) i tried it again from the console terminal, and guess what, it worked !! Now i cant recheck, but maybe this failure happens only if you are connected through an terminal-session with your server !?!? Ciao, Mario ---------------------------(end of broadcast)--------------------------- TIP 1: subscribe and unsubscribe commands go to majordomo@postgresql.org
> -----Original Message----- > From: pgsql-hackers-win32-owner@postgresql.org > [mailto:pgsql-hackers-win32-owner@postgresql.org] On Behalf > Of Rodrigo Moreno > Sent: 19 August 2004 18:54 > To: Mario Ivankovits; pgsql-hackers-win32@postgresql.org > Subject: RES: [pgsql-hackers-win32] InitDB Failure on install > > Yep, > > I detect this too. The problem happens in terminal services session. I'll try to add a check for this to the installer for the next release. It'll just be a simple 'you can't install this in a ts session' message though, not a fix. Mind you, it's certainly not the first installation I've seen that won't run in a ts session. Regards, Dave.
Dave Page wrote: > > > >>-----Original Message----- >>From: pgsql-hackers-win32-owner@postgresql.org >>[mailto:pgsql-hackers-win32-owner@postgresql.org] On Behalf >>Of Rodrigo Moreno >>Sent: 19 August 2004 18:54 >>To: Mario Ivankovits; pgsql-hackers-win32@postgresql.org >>Subject: RES: [pgsql-hackers-win32] InitDB Failure on install >> >>Yep, >> >>I detect this too. The problem happens in terminal services session. > > > I'll try to add a check for this to the installer for the next release. > It'll just be a simple 'you can't install this in a ts session' message > though, not a fix. Mind you, it's certainly not the first installation > I've seen that won't run in a ts session. Maybe it works if ts is put in install mode before starting the install? Regards, Andreas
> -----Original Message----- > From: Andreas Pflug [mailto:pgadmin@pse-consulting.de] > Sent: 19 August 2004 21:13 > To: Dave Page > Cc: Rodrigo Moreno; Mario Ivankovits; > pgsql-hackers-win32@postgresql.org > Subject: Re: [pgsql-hackers-win32] InitDB Failure on install > > Dave Page wrote: > > > > > > > >>-----Original Message----- > >>From: pgsql-hackers-win32-owner@postgresql.org > >>[mailto:pgsql-hackers-win32-owner@postgresql.org] On Behalf > Of Rodrigo > >>Moreno > >>Sent: 19 August 2004 18:54 > >>To: Mario Ivankovits; pgsql-hackers-win32@postgresql.org > >>Subject: RES: [pgsql-hackers-win32] InitDB Failure on install > >> > >>Yep, > >> > >>I detect this too. The problem happens in terminal services session. > > > > > > I'll try to add a check for this to the installer for the > next release. > > It'll just be a simple 'you can't install this in a ts session' > > message though, not a fix. Mind you, it's certainly not the first > > installation I've seen that won't run in a ts session. > > Maybe it works if ts is put in install mode before starting > the install? Possibly, but other apps which I've seen refuse to install on TS sessions have done so with TS in administrative mode in which there is no install mode. Can anyone test this? I can't start playing with our TS boxes :-( Regards, Dave
To add a little to the confusion, this does not seem to occur using 2003 terminal server. I have been building from source and running on this setup for the last few months (actually been using rdesktop from a Linux workstation -> win 2003 ). Note that I have not tried to install by means of the installer, and also have not tried to build the latest beta (last build was about 2.5 weeks ago). regards Mark Mario Ivankovits wrote: >> FATAL: could not create shared memory segment: No error >> DETAIL: Failed system call was shmget(key=1, size=1196032, 03600). >> child process was terminated by signal 1 >> initdb: failed >> initdb: removing cont > > > I have had the same problem if i tried to install postgresql within an > terminal-session on your windows 2000 terminal-server. > > Today (after a reboot) i tried it again from the console terminal, and > guess what, it worked !! > > Now i cant recheck, but maybe this failure happens only if you are > connected through an terminal-session with your server !?!? > > > Ciao, > Mario > > > ---------------------------(end of broadcast)--------------------------- > TIP 1: subscribe and unsubscribe commands go to majordomo@postgresql.org
Andreas Pflug wrote: >> I'll try to add a check for this to the installer for the next release. >> It'll just be a simple 'you can't install this in a ts session' message >> though, not a fix. Mind you, it's certainly not the first installation >> I've seen that won't run in a ts session. > Should be sufficient. > Maybe it works if ts is put in install mode before starting the install? I removed our postgres installation and tried to install it in "install mode" through an ts session - it doesnt worked anyway. So i went to our console and reinstalled it - with success. -- Mario
Dave Page schrieb:
It cost me a couple of hours to figure out that neither hibernate nor I have done something wrong when i tried to update a record with a timestamp in its key.
I know, these "couple of hours" stands in no relation to what you guys spent to develop and support such a _great_ database. However a short:
"Automatic timezone detection failed. Please edit your postgresql.conf (timezone=) manually to reflect your timezone. This will be fixed in the release version of PostgreSQL 8.0" ;-)
would be nice.
<>I'll try to add a check for this to the installer for the next release.I dont know what the state of the "timzone detection problem" in the current postgresql 8.0 head will be, but if it isnt fixed for the next beta it might greatly help others if you could add a warning to the installer if the automatic timezone detection failed.
It'll just be a simple 'you can't install this in a ts session' message
though, not a fix. Mind you, it's certainly not the first installation
I've seen that won't run in a ts session.
It cost me a couple of hours to figure out that neither hibernate nor I have done something wrong when i tried to update a record with a timestamp in its key.
I know, these "couple of hours" stands in no relation to what you guys spent to develop and support such a _great_ database. However a short:
"Automatic timezone detection failed. Please edit your postgresql.conf (timezone=) manually to reflect your timezone. This will be fixed in the release version of PostgreSQL 8.0" ;-)
would be nice.
-- Mario
> -----Original Message----- > From: pgsql-hackers-win32-owner@postgresql.org > [mailto:pgsql-hackers-win32-owner@postgresql.org] On Behalf > Of Mario Ivankovits > Sent: 20 August 2004 07:40 > To: pgsql-hackers-win32@postgresql.org > Cc: Andreas Pflug > Subject: Re: [pgsql-hackers-win32] InitDB Failure on install > > > I removed our postgres installation and tried to install it > in "install mode" through an ts session - it doesnt worked anyway. > > So i went to our console and reinstalled it - with success. OK, thanks. I committed a patch for this last night BTW. One thing that struck me this morning though - the patch checks that %SESSIONNAME% == "Console" (as oppsed to something like ICA-tcp#45 as you might see on a Citrix session); does the word "Console" get localised on non-English versions of Windows? Can someone check this for me please? Regards, Dave
Dave Page schrieb: >One thing that struck me this morning though - the patch checks that %SESSIONNAME% == >"Console" (as oppsed to something like ICA-tcp#45 as you might see on a >Citrix session); does the word "Console" get localised on non-English >versions of Windows? Can someone check this for me please? > > At least here on our german installation the SESSIONNAME is also "Console". I have checked it now. -- Mario
> -----Original Message----- > From: Mario Ivankovits [mailto:imario@apache.org] > Sent: 20 August 2004 08:21 > To: Dave Page > Cc: pgsql-hackers-win32@postgresql.org > Subject: Re: [pgsql-hackers-win32] InitDB Failure on install > > Dave Page schrieb: > > >One thing that struck me this morning though - the patch checks that > >%SESSIONNAME% == "Console" (as oppsed to something like > ICA-tcp#45 as > >you might see on a Citrix session); does the word "Console" get > >localised on non-English versions of Windows? Can someone > check this for me please? > > > > > At least here on our german installation the SESSIONNAME is > also "Console". I have checked it now. Thanks. Regards, Dave
> > I removed our postgres installation and tried to install it in > > "install mode" through an ts session - it doesnt worked anyway. > > > > So i went to our console and reinstalled it - with success. > > OK, thanks. I committed a patch for this last night BTW. One > thing that struck me this morning though - the patch checks > that %SESSIONNAME% == "Console" (as oppsed to something like > ICA-tcp#45 as you might see on a Citrix session); does the > word "Console" get localised on non-English versions of > Windows? Can someone check this for me please? Um. Is this really a problem with the installer? IMHO it seems more like a problem with the backend, so if we manage to fix that one be prepared to remoev the check :-) //Magnus
> -----Original Message----- > From: Magnus Hagander [mailto:mha@sollentuna.net] > Sent: 20 August 2004 12:00 > To: Dave Page; Mario Ivankovits; pgsql-hackers-win32@postgresql.org > Cc: Andreas Pflug > Subject: RE: [pgsql-hackers-win32] InitDB Failure on install > > > > I removed our postgres installation and tried to install it in > > > "install mode" through an ts session - it doesnt worked anyway. > > > > > > So i went to our console and reinstalled it - with success. > > > > OK, thanks. I committed a patch for this last night BTW. One thing > > that struck me this morning though - the patch checks that > > %SESSIONNAME% == "Console" (as oppsed to something like > > ICA-tcp#45 as you might see on a Citrix session); does the word > > "Console" get localised on non-English versions of Windows? Can > > someone check this for me please? > > Um. Is this really a problem with the installer? IMHO it > seems more like a problem with the backend, so if we manage > to fix that one be prepared to remoev the check :-) I've got no problem with that - but we've seen a number of reports of this, and no hint that anyone has the foggiest why it's happening. At least this will stave off a more reports next time round. Regards, Dave.
> > > > I removed our postgres installation and tried to install it in > > > > "install mode" through an ts session - it doesnt worked anyway. > > > > > > > > So i went to our console and reinstalled it - with success. > > > > > > OK, thanks. I committed a patch for this last night BTW. > One thing > > > that struck me this morning though - the patch checks that > > > %SESSIONNAME% == "Console" (as oppsed to something like > > > ICA-tcp#45 as you might see on a Citrix session); does the word > > > "Console" get localised on non-English versions of Windows? Can > > > someone check this for me please? > > > > Um. Is this really a problem with the installer? IMHO it seems more > > like a problem with the backend, so if we manage to fix that one be > > prepared to remoev the check :-) > > I've got no problem with that - but we've seen a number of > reports of this, and no hint that anyone has the foggiest why > it's happening. At least this will stave off a more reports > next time round. I'm working on the issue. First I have to fix the error messages though, but I think I clearly have enough to reproduce it consitently now at least. //Magnus
> -----Original Message-----
> From: Andreas Pflug [mailto:pgadmin@pse-consulting.de]
> Sent: 20 August 2004 13:21
> To: Dave Page
> Cc: pgsql-hackers-win32@postgresql.org
> Subject: Re: [pgsql-hackers-win32] InitDB Failure on install
>
> Dave Page wrote:
>
> >
> >
> > OK, thanks. I committed a patch for this last night BTW. One thing
> > that struck me this morning though - the patch checks that
> > %SESSIONNAME% == "Console" (as oppsed to something like
> ICA-tcp#45 as
> > you might see on a Citrix session); does the word "Console" get
> > localised on non-English versions of Windows? Can someone
> check this for me please?
>
> The official way to detect a terminal service session
> according to MSDN is
>
> bool IsRemoteSession() { return GetSystemMetrics(SM_REMOTESESSION); }
Hmm, too much time playing with Citrix recently! I'll change it shortly
- thanks.
Regards, Dave
> -----Original Message-----
> From: pgsql-hackers-win32-owner@postgresql.org
> [mailto:pgsql-hackers-win32-owner@postgresql.org] On Behalf
> Of Dave Page
> Sent: 20 August 2004 12:08
> To: Andreas Pflug
> Cc: pgsql-hackers-win32@postgresql.org
> Subject: Re: [pgsql-hackers-win32] InitDB Failure on install
>
> > The official way to detect a terminal service session according to
> > MSDN is
> >
> > bool IsRemoteSession() { return
> GetSystemMetrics(SM_REMOTESESSION); }
>
> Hmm, too much time playing with Citrix recently! I'll change
> it shortly
> - thanks.
Actually I'm going to leave it as it is for now - that won't work on NT4
(even with TS afaict), and as it's possible the installer may be fixed
for NT in the future, I'd rather not introduce any further
incompatibilities.
Regards, Dave.
> -----Original Message----- > From: Andreas Pflug [mailto:pgadmin@pse-consulting.de] > Sent: 20 August 2004 14:30 > To: Dave Page > Cc: pgsql-hackers-win32@postgresql.org > Subject: Re: [pgsql-hackers-win32] InitDB Failure on install > > No, it's NT4SP4 or later. > We definitely should require NT4SP4 or later anyway. Not according to MSDN (http://msdn.microsoft.com/library/default.asp?url=/library/en-us/termse rv/termserv/detecting_the_terminal_services_environment.asp): The following code sample shows a function that returns TRUE if the application is running in a remote session and FALSE if the application is running on the console. It works only on Windows 2000 and later. > BTW, GetSystemMetrics would return 0 for unsupported SM_ values. True. Regards, Dave.
Dave Page wrote:
>
>
> OK, thanks. I committed a patch for this last night BTW. One thing that
> struck me this morning though - the patch checks that %SESSIONNAME% ==
> "Console" (as oppsed to something like ICA-tcp#45 as you might see on a
> Citrix session); does the word "Console" get localised on non-English
> versions of Windows? Can someone check this for me please?
The official way to detect a terminal service session according to MSDN is
bool IsRemoteSession() { return GetSystemMetrics(SM_REMOTESESSION); }
Regards,
Andreas
Dave Page wrote:
>
>
>
>>-----Original Message-----
>>From: pgsql-hackers-win32-owner@postgresql.org
>>[mailto:pgsql-hackers-win32-owner@postgresql.org] On Behalf
>>Of Dave Page
>>Sent: 20 August 2004 12:08
>>To: Andreas Pflug
>>Cc: pgsql-hackers-win32@postgresql.org
>>Subject: Re: [pgsql-hackers-win32] InitDB Failure on install
>>
>>
>>>The official way to detect a terminal service session according to
>>>MSDN is
>>>
>>>bool IsRemoteSession() { return
>>
>>GetSystemMetrics(SM_REMOTESESSION); }
>>
>>Hmm, too much time playing with Citrix recently! I'll change
>>it shortly
>>- thanks.
>
>
> Actually I'm going to leave it as it is for now - that won't work on NT4
> (even with TS afaict), and as it's possible the installer may be fixed
> for NT in the future, I'd rather not introduce any further
> incompatibilities.
No, it's NT4SP4 or later.
We definitely should require NT4SP4 or later anyway.
BTW, GetSystemMetrics would return 0 for unsupported SM_ values.
Regards,
Andreas