From: | Kris Kiger <kris(at)musicrebellion(dot)com> |
---|---|
To: | Jeff Frost <jeff(at)frostconsultingllc(dot)com> |
Cc: | pgsql-admin(at)postgresql(dot)org |
Subject: | Re: archive_command |
Date: | 2005-09-28 17:08:18 |
Message-ID: | 433ACE02.3040208@musicrebellion.com |
Views: | Raw Message | Whole Thread | Download mbox | Resend email |
Thread: | |
Lists: | pgsql-admin |
Jeff: I see, that does make a lot of sense ;-). Thanks
However, that answer brings me to another question:
To begin with an example, lets say we start the server and are writing
to wal file X, however postgres has preallocated wal files up to X+7.
We assume that archive command will be invoked on wal files as they
become full. We don't care about wal files that are greater than X,
because they have either been recycled or preallocated, in either case
the data has been backed up from a previous archive command or does not
yet exist. So, we can assume that all relevant wal logs have been
archived, save the current one.
My question then is, how do we tell which wal log is X (the one
currently being written to)?
Jeff Frost wrote:
> Kris,
>
> It is invoked when the WAL file is filled. I believe the server
> preallocates 8 or so of them on startup.
>
> On Wed, 28 Sep 2005, Kris Kiger wrote:
>
>> set to use rsync copy the wal log from server A to server B. When
>> i'm watching the number of wal files on server A I notice sometimes
>> there are 3 files that have not been archived, sometimes there are 4,
>> and sometimes there are more. So, my question is, when is
>> archive_command invoked?
>
>
From | Date | Subject | |
---|---|---|---|
Next Message | Jeff Frost | 2005-09-28 17:15:20 | Re: archive_command |
Previous Message | jeff shu | 2005-09-28 17:05:18 | db connection and script execution error |