Resources available, website down
I have run into a problem where the website stays down even when there are plenty of resources available. It is currently at about one third of its usual traffic, and somehow keeps going down at random.
I'm using only a small number of plugins (8 active). No cron jobs. And there's plenty of RAM, hard disk space and CPU to go around.
No major changes have been made to the website in a while, except the disabling of a plugin (jetpack).
Any help would be greatly appreciated. Thank you.
3 Replies
Pingdom found it to be running the whole time. My computers could rarely access it, downforeveryoneorjustme agreed with my computers, so do access reports, web counters and logs.
Nothing specific shows up in the error log that would match the outages.
> 130903 8:04:46 [Note] Event Scheduler: Purging the queue. 0 events
130903 8:04:46 InnoDB: Starting shutdown…
130903 8:11:41 [Note] Plugin 'FEDERATED' is disabled.
130903 8:11:41 InnoDB: Initializing buffer pool, size = 8.0M
130903 8:11:41 InnoDB: Completed initialization of buffer pool
130903 8:11:42 InnoDB: Started; log sequence number 0 44233
130903 8:11:42 [Note] Event Scheduler: Loaded 0 events
130903 8:11:42 [Note] /usr/sbin/mysqld: ready for connections.
This is the last entry in the sql error log. That's about the time I issued a reboot.
It continued having problems after this, and seems to be automagically back to work for the last 40 minutes or so.
If it's Apache (in other words, if you can't serve up anything whether it's a Wordpress site or a single html page with no server-side processing), possibly a SYN flood? Next time your site becomes unresponsive, go into a shell, run "netstat -nt" and see if you have tens/hundreds/thousands of lines that say "SYN RECV" under the "State" column. If you're not set up to mitigate it, this will bring Apache to its knees while the rest of your box hums away like nothing is wrong, and IIRC nothing will show in the logs.