Topic: Weird production Logs

Some of the snippets from my production log. (Nginx+mongrel on a shared hosting server)

Completed in 222.34187 (0 reqs/sec) | Rendering: 3.91175 (1%) | DB: 133.66051 (60%) | 200 OK
Completed in 73.34449 (0 reqs/sec) | Rendering: 0.82197 (1%) | DB: 2.60048 (3%) | 200 OK
Completed in 332.15385 (0 reqs/sec) | Rendering: 5.15014 (1%) | DB: 10.90407 (3%) | 200 OK

This mostly happens if there are long idle times (o/w site works fine). So, I tried changing the verification_timeout in config file but that doesn't seems to solve the problem. (I have pretty small database of size 4k properly indexed.)

Any ideas what might the problem be?
Thanks in advance.

Re: Weird production Logs

It would be interesting to see what top and vmstat show at the time...  (I'm assuming linux here.)  Is it possible the servers are massively overloaded?

BrewControl.com - Brewery and Brewpub management powered by RoR

Re: Weird production Logs

Currently its working fine but last time when it happened and I checked,  nothing was changing for mongrel+nginx processes. (both cpu & memory consumption was constant to ~0.1)

Support guys says overloading is not the issue. No one else on that shared server is facing this problem. "mysql slow query log" doesn't reports any problems either.

Last edited by nakula (2007-04-17 06:34:15)