| From: | Thomas Munro <thomas(dot)munro(at)gmail(dot)com> |
|---|---|
| To: | Shawn Debnath <sdn(at)amazon(dot)com> |
| Cc: | Andres Freund <andres(at)anarazel(dot)de>, Robert Haas <robertmhaas(at)gmail(dot)com>, Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us>, Dmitry Dolgov <9erthalion6(at)gmail(dot)com>, PostgreSQL-development <pgsql-hackers(at)postgresql(dot)org> |
| Subject: | Re: Refactoring the checkpointer's fsync request queue |
| Date: | 2019-04-05 03:19:54 |
| Message-ID: | CA+hUKGL+YLUOA0eYiBXBfwW+bH5kFgh94=gQH0jHEJ-t5Y91wQ@mail.gmail.com |
| Views: | Raw Message | Whole Thread | Download mbox | Resend email |
| Thread: | |
| Lists: | pgsql-hackers |
On Thu, Apr 4, 2019 at 11:47 PM Thomas Munro <thomas(dot)munro(at)gmail(dot)com> wrote:
> Pushed. Thanks for all the work on this!
I managed to break this today while testing with RELSEG_SIZE set to 1
block (= squillions of 8kb files). The problem is incorrect arguments
to _mdfd_getseg(), in code added recently by me. Without the
EXTENSION_DONT_CHECK_SIZE flag, it refuses to open segments following
segments that have been truncated, leading to a checkpointer fsync
panic. It's also passing segnum where a blocknum is wanted. It
should have used exactly the same arguments as in the old code, but
didn't. I will push a fix shortly.
--
Thomas Munro
https://enterprisedb.com
| From | Date | Subject | |
|---|---|---|---|
| Next Message | Amit Langote | 2019-04-05 03:25:15 | Re: speeding up planning with partitions |
| Previous Message | David Rowley | 2019-04-05 03:18:41 | Re: speeding up planning with partitions |