Re: [9.3 bug] disk space in pg_xlog increases during archive recovery - Mailing list pgsql-hackers

From Fujii Masao
Subject Re: [9.3 bug] disk space in pg_xlog increases during archive recovery
Date
Msg-id CAHGQGwEUGpTL6jNObjXTKp_JvBgF0DGSC4v9Wy+eymfh2Fmo5Q@mail.gmail.com
Whole thread Raw
In response to Re: [9.3 bug] disk space in pg_xlog increases during archive recovery  ("MauMau" <maumau307@gmail.com>)
Responses Re: [9.3 bug] disk space in pg_xlog increases during archive recovery
List pgsql-hackers
On Wed, Jul 31, 2013 at 10:43 PM, MauMau <maumau307@gmail.com> wrote:
> Hello, Fujii san, all,
>
> From: "Fujii Masao" <masao.fujii@gmail.com>
>>
>> On Sun, Jul 28, 2013 at 7:59 AM, MauMau <maumau307@gmail.com> wrote:
>> Do you think this should be fixed?
>>
>> I think so.
>>
>>> How should it be fixed?
>>
>>
>> What about removing the restored archived file as soon as it's replayed
>> if cascading replication is not enabled (i.e., max_wal_senders = 0 or
>> hot_standby = off)? This doesn't seem to break the existing behavior
>> in 9.2.
>
>
> Please find attached the patch to fix the problem.  I changed to keep
> restored WAL files in pg_xlog/ only on a cascading standby server.  Could
> you review and commit this?

-    if (source == XLOG_FROM_ARCHIVE)
+    if (source == XLOG_FROM_ARCHIVE &&
+        StandbyModeRequested && AllowCascadeReplication())

I think that the condition of StandbyModeRequested should be removed
because someone might want to set up the cascade standby from the standby
of warm-standby configuration.

> BTW, KeepFileRestoredFromArchive() is also called to keep timeline history
> files in pg_xlog/.  What is this for?  Is this necessary for recovery other
> than cascading standbys?

Yes. Please see 60df192aea0e6458f20301546e11f7673c102101

> This seems to accumulate timeline history files
> forever in pg_xlog/.

Yes, because basically there is no way to delete the timeline history files
from pg_xlog.

Regards,

-- 
Fujii Masao



pgsql-hackers by date:

Previous
From: Tom Lane
Date:
Subject: Re: REGEXP_MATCHES() strange behavior with '^' and '$' pattern
Next
From: Greg Stark
Date:
Subject: Re: Review: UNNEST (and other functions) WITH ORDINALITY