InnoDB hang when run sysbench test

[louis...@gmail.com]

Hi, all ,

When I run sysbench test for mysql, i encounter the following Warnings:

Version: '5.5.43-37.2-log'  socket: '/home/q/mysql/multi/3306/socket/mysql.sock'  port: 3306  Percona Server (GPL)

InnoDB: Warning: a long semaphore wait:

--Thread k140317364688640 has waited at btr0cur.c line 570 for 241.00 seconds the semaphore:

X-lock (wait_ex) on RW-latch at 0x7f9e28020228 '&new_index->lock'

a writer (thread id 140317364688640) has reserved it in mode  wait exclusive

number of readers 1, waiters flag 1, lock_word: ffffffffffffffff

All threads hang for something!!!!!  

I found https://bugs.mysql.com/bug.php?id=74743 and http://bugs.mysql.com/bug.php?id=73890

But i found there is a little difference with my SEMAPHORES warnings.

The log can be found at http://pastebin.com/6epg3NtG

[louis...@gmail.com]

[MARK CALLAGHAN]

One possible cause is unhealthy storage -- a disk read or write might be stuck. Thread stacks obtained via PMP (http://poormansprofiler.org) might help to explain this.

--
You received this message because you are subscribed to the Google Groups "Percona Discussion" group.
To unsubscribe from this group and stop receiving emails from it, send an email to percona-discuss...@googlegroups.com.
To post to this group, send email to percona-d...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

--

Mark Callaghan
mdca...@gmail.com

[Justin Swanhart]

Hi,

Checking dmesg and syslog never hurt, nor does using mcelog to check for machine exception events such as overheating processors, bad ECC memory etc.

Long semaphore waits can sometimes be caused by bugs, but from experience, almost all are caused by some misbehaving piece of hardware or software, such as a bad RAID controller.  If you have a hardware RAID controller, check to see if it was performing a test and had disabled the write-back cache in favor of write-through during the test.

Regards,

[MARK CALLAGHAN]

If you can't get output from PMP, then "SHOW ENGINE INNODB STATUS" is the next best thing