| From: | Asim Praveen <apraveen(at)pivotal(dot)io> |
|---|---|
| To: | Michael Paquier <michael(dot)paquier(at)gmail(dot)com> |
| Cc: | Alvaro Herrera <alvherre(at)alvh(dot)no-ip(dot)org>, Jacob Champion <pchampion(at)pivotal(dot)io>, Robert Haas <robertmhaas(at)gmail(dot)com>, PostgreSQL mailing lists <pgsql-hackers(at)postgresql(dot)org>, Kevin Grittner <kgrittn(at)gmail(dot)com> |
| Subject: | Re: [PATCH] Assert that the correct locks are held when calling PageGetLSN() |
| Date: | 2017-11-08 23:25:32 |
| Message-ID: | CANXE4TfddjO1oBbQsX-CsBfG=n6+kE1hPEksyuviH8B6va5nTQ@mail.gmail.com |
| Views: | Raw Message | Whole Thread | Download mbox | Resend email |
| Thread: | |
| Lists: | pgsql-hackers |
Hi Michael
On Mon, Nov 6, 2017 at 6:18 PM, Michael Paquier <michael(dot)paquier(at)gmail(dot)com>
wrote:
>
> Did you really test WAL replay? This still ignores that PageGetLSN is
> as well taken in some code paths, like recovery, where actions on the
> page are guaranteed to be serialized, like during recovery, so this
> patch would cause the system to blow up. Note that pageinspect,
> amcheck and wal_consistency_checking also process on page copies. So
> the assertion failure of 0002 would trigger in those cases.
>
Indeed, the assertion tripped during WAL replay on the standby. This was
caught by TAP tests under src/test/recovery. The assertion is now fixed so
that WAL replay is exempt from the check. Please find the new patch
attached. The tests now pass with the fix. I also manually verified that
recovery works with "wal_consistency_checking=all".
Asim
| Attachment | Content-Type | Size |
|---|---|---|
| 0002-PageGetLSN-assert-that-locks-are-properly-held.patch | application/octet-stream | 3.9 KB |
| From | Date | Subject | |
|---|---|---|---|
| Next Message | Simon Riggs | 2017-11-08 23:48:38 | Re: Transaction control in procedures |
| Previous Message | Michael Paquier | 2017-11-08 23:19:52 | Re: taking stdbool.h into use |