List:General Discussion« Previous MessageNext Message »
From:mysql Date:August 4 2002 3:59am
Subject:SIGPROF problem in FreeBSD 4.6 (REPOST)
View as plain text  
I thought I got this through the list, but it doesn't show up in the
search, and I didn't get any responses...

---------- Forwarded message ----------
Date: Wed, 31 Jul 2002 23:14:13 -0400 (EDT)
From: mysql@stripped
To: mysql@stripped
Subject: SIGPROF problem in FreeBSD 4.6

This almost belongs on the bugs list, but I don't have a stable, sure way
to trigger it.

Symptoms:

mysqld locks itself in a tight loop, consuming all available CPU cycles.
a simple "kill" sometimes kills the daemon, other times a -9 is required.
While in this state, any access to the daemon stops - mysql queries simply
hang.

ktrace -p <pid> -t+ shows this repeating endlessly:

44113 mysqld   PSIG  SIGPROF caught handler=0x812d58c mask=0x0 code=0x0
 44113 mysqld   CALL  gettimeofday(0x81c634c,0)
 44113 mysqld   RET   gettimeofday 0
 44113 mysqld   CALL  sigprocmask(0x3,0x81c63d8,0)
 44113 mysqld   RET   sigprocmask 0
 44113 mysqld   CALL  sigaltstack(0x8203ca0,0)
 44113 mysqld   RET   sigaltstack 0
 44113 mysqld   CALL  poll(0x821c000,0x2,0)
 44113 mysqld   RET   poll 0
 44113 mysqld   CALL  sigreturn(0x24473064)
 44113 mysqld   RET   sigreturn JUSTRETURN
 44113 mysqld   PSIG  SIGPROF caught handler=0x812d58c mask=0x0 code=0x0
 44113 mysqld   CALL  gettimeofday(0x81c634c,0)
 44113 mysqld   RET   gettimeofday 0
 44113 mysqld   CALL  sigprocmask(0x3,0x81c63d8,0)
 44113 mysqld   RET   sigprocmask 0
 44113 mysqld   CALL  sigaltstack(0x8203ca0,0)
 44113 mysqld   RET   sigaltstack 0
 44113 mysqld   CALL  poll(0x821c000,0x2,0)
 44113 mysqld   RET   poll 0
 44113 mysqld   CALL  sigreturn(0x24473064)
 44113 mysqld   RET   sigreturn JUSTRETURN
 44113 mysqld   PSIG  SIGPROF caught handler=0x812d58c mask=0x0 code=0x0
 44113 mysqld   CALL  gettimeofday(0x81c634c,0)
 44113 mysqld   RET   gettimeofday 0
 44113 mysqld   CALL  sigprocmask(0x3,0x81c63d8,0)
 44113 mysqld   RET   sigprocmask 0
 44113 mysqld   CALL  sigaltstack(0x8203ca0,0)
 44113 mysqld   RET   sigaltstack 0
 44113 mysqld   CALL  poll(0x821c000,0x2,0)
 44113 mysqld   RET   poll 0

It's happening fairly often now.

It's not load related - nothing besides a simple data loading program is
running when it enters this loop.

It's not seemingly data related:  I've run myisamchk --safe-recover on the
entire database and had the problem an hour later.  I've started loading
the data over and had the problem not recur.

It's not strictly version related: I've had the same problem with 3.23.51,
4.0.1-alpha, and 4.0.2-alpha.

The program that causes it is a simple perl script that reads input data,
splits it into fields, validates the data, and uses a REPLACE statement
with DBD to update/insert the table.

One thing that may be related is that there are quite a few rows in the
table in question - something over 197 million now.  I don't recall the
problem happening at all with fewer than 180 to 190 million rows.

I'd love to track this down.  I could use some help.  Suggestions on what
I can  do to help are most welcome.

Thanks,

--- David





Thread
SIGPROF problem in FreeBSD 4.6 (REPOST)mysql4 Aug
  • Re: SIGPROF problem in FreeBSD 4.6 (REPOST)Jeremy Zawodny4 Aug
    • Re: SIGPROF problem in FreeBSD 4.6 (REPOST)David Miller10 Aug
      • Re: SIGPROF problem in FreeBSD 4.6 (REPOST)Dan Nelson10 Aug
        • Re: SIGPROF problem in FreeBSD 4.6 (REPOST)Jeremy Zawodny10 Aug
      • Re: SIGPROF problem in FreeBSD 4.6 (REPOST)Jeremy Zawodny18 Aug