On Mon, 15 Apr 2024 at 14:55, David Rowley <
dgrowleyml@gmail.com> wrote:
> If GetNowFloat() somehow was returning a negative number then we could
> end up with a large delay. But if gettimeofday() was so badly broken
> then wouldn't there be some evidence of this in the log timestamps on
> failing runs?
3 things stand out for me here, unsure if they're related somehow:
1. Issue where reltuples=48 (in essence runs complete, but few tests fail)
2. SIGABRT - most of which are DDLs (runs complete, but engine crashes + many tests fail)
3. pg_sleep() stuck - (runs never complete, IIUC never gets reported to buildfarm)
For #3, one thing I had done earlier (and then reverted) was to set the
'wait_timeout' from current undef to 2 hours. I'll set it again to 2hrs
in hopes that #3 starts getting reported to buildfarm too.
> I'm not that familiar with the buildfarm config, but I do see some
> Valgrind related setting in there. Is PostgreSQL running under
> Valgrind on these runs?
Not yet. I was tempted, but valgrind has not yet been enabled on
this member. IIUC by default they're disabled.
-
robins