Bug #4843 Segfault when server closes down
Submitted: 31 Jul 2004 22:02 Modified: 1 Aug 2004 10:11
Reporter: [ name withheld ] Email Updates:
Status: Closed Impact on me:
None 
Category:MySQL Administrator Severity:S3 (Non-critical)
Version:1.0.8a OS:Linux (Debian/Linux)
Assigned to: CPU Architecture:Any

[31 Jul 2004 22:02] [ name withheld ]
Description:
Segfault. This could be a bug in mysql client library. (MySQL 4.0.20) Leaving it as priority low/non-critical since it only occurs under specific circumstances.

Partial backtrace,

#0  0x40312451 in __write_nocancel () from /lib/tls/libpthread.so.0
#1  0x40047f6f in vio_write () from /usr/lib/libmysqlclient.so.12
#2  0x40049c88 in net_real_write () from /usr/lib/libmysqlclient.so.12
#3  0x40049778 in net_flush () from /usr/lib/libmysqlclient.so.12
#4  0x400499e0 in net_write_command () from /usr/lib/libmysqlclient.so.12
#5  0x40030257 in simple_command () from /usr/lib/libmysqlclient.so.12
#6  0x4003396a in mysql_ping () from /usr/lib/libmysqlclient.so.12
#7  0x08090a9c in MInstanceInfo::check_connection (this=0xbffff4a0, mysql=0x40062e24) at MInstanceInfo.cc:302
#8  0x0809161d in MInstanceInfo::perform_data_fetch (this=0xbffff4a0, fetcher=0xffffffe0, msg=@0xbfffe660)
    at MInstanceInfo.cc:505
#9  0x080d5c60 in MAHealthPanel::refresh_list (this=0x82498d0, index=-32) at MAHealthPanel.cc:1070

How to repeat:
1. Open MySQL Admin and connect to a server (don't do anything else here)
2. Shut down the server (using a command line)
3. In MySQL Admin, click on Health
[1 Aug 2004 10:11] Alfredo Kojima
I've added checks to see if the connection was lost before sending requests to the server and made the code a bit more robust. It should be more stable when the server goes down, now.