Re: commit dfda6ebaec67 versus wal_keep_segments

On 03.04.2013 18:58, Jeff Janes wrote:
> On Tue, Apr 2, 2013 at 10:08 PM, Jeff Janes<jeff(dot)janes(at)gmail(dot)com> wrote:
>
>> This commit introduced a problem with wal_keep_segments:
>>
>> commit dfda6ebaec6763090fb78b458a979b558c50b39b
>
> The problem seems to be that the underflow warned about is happening,
> because the check to guard it was checking the wrong thing. However, I
> don't really understand KeepLogSeg. It seems like segno, and hence recptr,
> don't actually serve any purpose.

Hmm, the check is actually correct, but the assignment in the
else-branch isn't. The idea of KeepLogSeg is to calculate recptr -
wal_keep_segments, and assign that to *logSegNo. But only if *logSegNo
is not already < than the calculated value. Does the attached look
correct to you?

> At some point when it is over-pruning and recycling, it recyles the log
> files that are still needed for recovery, and if the database crashes at
> that point it will not recover because it can't find either the primary
> secondary checkpoint records.

So, KeepLogSeg incorrectly sets *logSegNo to 0, and CreateCheckPoint
decrements it, causing it to underflow to 2^64-1. Now RemoveOldXlogFiles
feels free to remove every WAL segment.

- Heikki