reaper dies if DB connection closes

GoogleCodeExporter commented 8 years ago

What steps will reproduce the problem?
1. Configure MySQL with low wait_timeout value (300)
2. Mark a device as dead
3. Wait until all rows have been removed from file_on

What is the expected output? What do you see instead?

No output. Reaper periodically checks the file_on table, and eventually forgets 
about the device.

Because reaper sleeps longer than "wait_timeout" MySQL closes the connection 
and reaper dies with:

mogilefsd[26521]: crash log: DBD::mysql::db selectrow_array failed: MySQL 
server has gone away at /usr/local/share/perl/5.10.0/MogileFS/Store/MySQL.pm 
line 114.

What version of the product are you using? On what operating system?

2.65

Original issue reported on code.google.com by pyh...@gmail.com on 1 Nov 2012 at 7:54

GoogleCodeExporter commented 8 years ago

I was having the same issue, with the only difference that I have haproxy in 
front of two master-master mysql servers running the mogilefs database. The 
client and server timeout values were set to 50s and I was seeing those crash 
messages every 77 seconds.

After the timeouts were increased to 500s, the errors disappeared.

Original comment by yout...@shapirus.net on 11 Dec 2012 at 7:00

GoogleCodeExporter commented 8 years ago

This has been fixed in 2.66.

Original comment by dorma...@rydia.net on 7 Jan 2013 at 5:41

Changed state: Fixed

glyustb / mogilefs

reaper dies if DB connection closes #75