stats collector is stuck on Windows - Mailing list pgsql-hackers

From ITAGAKI Takahiro
Subject stats collector is stuck on Windows
Date
Msg-id 20060905113341.5A96.ITAGAKI.TAKAHIRO@oss.ntt.co.jp
Whole thread Raw
List pgsql-hackers
Hi Hackers,

I saw stats information was not updated sometimes when I tested HEAD on
Windows. Once it happens, stats are not collected until server restarts.

The condition seems to happen on high-CPU load. I can reproduce it
in a few seconds using pgbench on the machine with one Pentium 4
with hyper-threading. But if I use an original "pgbench with sleep"
load, it won't occur.

Is this a bug or some environment-dependent problems?



[additional information]
I added elog calls for printf-debug.
 PgstatCollectorMain() in pgstat.c | elog(LOG, "pgstat select"); | if (select(pgStatSock + 1, &rfds, NULL, NULL,
&sel_timeout)< 0) | { ...
 

After stuck, select() always results in time-out. "LOG:  pgstat select" is
repeated every 2 seconds (maybe PGSTAT_SELECT_TIMEOUT).

Regards,
---
ITAGAKI Takahiro
NTT Open Source Software Center



pgsql-hackers by date:

Previous
From: Bruce Momjian
Date:
Subject: Re: Sorry about the GIN docs :(
Next
From: Alvaro Herrera
Date:
Subject: Re: Open items for 8.2