After Update to SVN 4928 web pages time out constantly

[Joseph Hales]

The web page seems to be working now but I am still having some issues and the Apache stops logging a min after restart with the same errors as shown in the attached log.
Apache.log.txt

[Wayne Workman]

@Joseph-Hales I wonder if the sheer amount of errors being recorded simply crash the logging mechanism… ?

There are levels of logging in daemons, you can adjust the level on a scale, some have different modes.

This was written for Ubuntu, but I would imagine that the things in here can be applied to any distro very similarly:
https://www.digitalocean.com/community/tutorials/how-to-configure-logging-and-log-rotation-in-apache-on-an-ubuntu-vps

It might be beneficial to turn up the verbosity of logging just to figure out what’s going on… and it might be beneficial to turn the verbosity down to lower cpu load … and so logging doesn’t crash.

[Tom Elliott]

I’m not 100% sure how much I like approaching things in this manner, but i’ve now made it so FOG will die if the db connection cannot happen. This should stop spamming logs with “mysqli::query() couldn’t fetch mysqli” error messages. It also “deadens” the page right then and there. This way resources aren’t trying to get data as in the “main page” as used to be the case.

This isn’t going to fix the issue here, but should help in lessening errors being thrown all over the place.

[Wayne Workman]

@Tom-Elliott said:

i’ve now made it so FOG will die if the db connection cannot happen.

Can you elaborate on that? Will it try to come alive later on? If the DB connection is broke, can fog try to restart mysql? Can there be a limit on how often it gets restarted? Sounds like it’s living dangerously.

[Joseph Hales]

As of right now fog has been running for 18 hours when I run

show full processlist;

I get 191 lines of

| 43572 | root | localhost:35810 | fog | Sleep | 3 | | NULL

at least I’m not seeing the locked delete entries any more. I think I have a old deleted snapin associated with some hosts what would be a command to select and delete it from my sql it may be causing the apache errors I am getting now.

Apache.log.txt

[Wayne Workman]

@Joseph-Hales Not sure… I’d have to have my hands on it to build the right command for you.

Just work on building a select command. Once you can get the data you want gone selected, then just convert it to a delete command.

[Tom Elliott]

Is it safe to solve this thread now? The process list isn’t bad by itself as you’re most likely using persistent connections. Many of those processes are sleeping because of the persistence.

[Joseph Hales]

I’m worried about the sigterm termination and the max hosts error every time MySQL starts I will be switching hardware soon and if these are underlying database issues I would like to resolve them before migration.

[Joseph Hales]

moving this to solved as we are now on a much higher SVN and about to replace the server entirely.

[Tom Elliott]

@Joseph-Hales I agree with the plan to solve the thread but not for the reasons you describe, rather for the fact this was likely somewhat related to the issues found earlier.