Re: Reliable WAL file shipping over unreliable network
От | Mark Kirkwood |
---|---|
Тема | Re: Reliable WAL file shipping over unreliable network |
Дата | |
Msg-id | 0bff7871-1f88-65a6-ac77-4bbba8f89889@catalyst.net.nz обсуждение исходный текст |
Ответ на | Re: Reliable WAL file shipping over unreliable network (David Steele <david@pgmasters.net>) |
Ответы |
Re: Reliable WAL file shipping over unreliable network
|
Список | pgsql-admin |
On 07/03/18 08:31, David Steele wrote: > On 3/6/18 2:00 PM, Stephen Frost wrote: >> * Rui DeSousa (rui.desousa@icloud.com) wrote: >>>> On Mar 5, 2018, at 10:02 AM, Stephen Frost <sfrost@snowman.net> wrote: >>>> It doesn’t- but please don't encourage partial solutions which have ver clear issues. >>> Then problem is there are no good base utilities that is useful with archive_command; unless you’re writing directlyto an NFS mount or a tape library. Even Barman recommends rsync with the archive_command; if you are unable to usepg_receivexlog solution. There are countless Postgres documentation out there that recommends use of rsync with the archive_command. >>> >>> Here a solution that will fsync() file on the other end. >> I'd encourage you to consider submitting this to a commitfest to get >> feedback on it, as a proper patch to add such a utlity to PG and which >> we could use in our documentation instead of saying "cp", which is >> pretty terrible. > +1. I believe there's been talk of a pgcopy utility before but it never > went anywhere. I think it would be worth it just to have a decent > example in the docs. > +1. I think if there is a strong feeling that the current docs are suggesting things that shouldn't actually be done, then we should provide an alternative in the distribution itself. Re Pgbackrest (and other backup tools that various Postgres companies have written) - while these may be great (Pgbackrest looks pretty good for instance), they don't come with the Postgres package itself, and our docs suggest something else. Hence discussion threads like this one! The added (new) issue that there may be some unusual bug with rsync + quotas and ZFS on BSD (at least) gives a bit more weight to this idea. regards Mark
В списке pgsql-admin по дате отправления: