| From: | "Higuchi, Daisuke" <higuchi(dot)daisuke(at)jp(dot)fujitsu(dot)com> |
|---|---|
| To: | 'Laurenz Albe' <laurenz(dot)albe(at)cybertec(dot)at>, "Sakai, Teppei" <sakai(dot)teppei(at)jp(dot)fujitsu(dot)com>, "pgsql-hackers(at)postgresql(dot)org" <pgsql-hackers(at)postgresql(dot)org> |
| Subject: | RE: Problem during Windows service start |
| Date: | 2019-01-09 05:28:29 |
| Message-ID: | 1803D792815FC24D871C00D17AE95905D795C9@g01jpexmbkw24 |
| Views: | Raw Message | Whole Thread | Download mbox | Resend email |
| Thread: | |
| Lists: | pgsql-hackers |
Hi,
This thread is inactive, but I want to solve this problem.
I think this problem rarely occurs in 10 or later version because of commit [1]. Because "pg_ctl start -w" wait for only PID file creation. It means that timeout is not occurred even if crash recovery takes a lot of times.
However, 9.6 or earlier still wait for long time recovery complete.
> How do you propose to fix it?
I think there are two solutions.
One solution is that status of Windows Service should be changed to "SERVICE_RUNNING" even if timeout is occurred because of long time recovery. I attached the patch of this solution.
Another solution is to backport commit [1] to 9.6 or earlier version. However this fix change the content of PID file, so incompatible change, I think.
I would appreciate it if you give any comments.
[1] https://github.com/postgres/postgres/commit/f13ea95f9e473a43ee4e1baeb94daaf83535d37c
Regards,
Daisuke, Higuchi
| Attachment | Content-Type | Size |
|---|---|---|
| windows_service_bug_fix.patch | application/octet-stream | 1.2 KB |
| From | Date | Subject | |
|---|---|---|---|
| Next Message | Michael Paquier | 2019-01-09 05:40:23 | Re: Problem during Windows service start |
| Previous Message | Haribabu Kommi | 2019-01-09 04:56:20 | Re: New function pg_stat_statements_reset_query() to reset statistics of a specific query |