Timeline for mysql running over 200% CPU according to atop - pt-query-digest suggests slow running query is a lookup in wp_options for 'autoload' = yes
Current License: CC BY-SA 4.0
9 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| Jul 9 at 3:58 | review | Close votes | |||
| Aug 4 at 3:02 | |||||
| Jul 8 at 21:21 | answer | added | Phill Ash | timeline score: 0 | |
| Jul 7 at 22:36 | comment | added | Phill Ash | Looks like some kind of nefarious activities... I could see a large number of connections on port 80. 200 odd at a time when the site isn't doing anything... Sadly just as I went to check what they were my lil one woke up so I had to sort that and by the time I was back CPU was back to normal and all the connections were gone. I'll have to be faster off the mark tomorrow. | |
| Jul 7 at 19:47 | comment | added | Phill Ash | Sadly it appears that the problem has just moved to php-fpm. Now I see mysqld using 200% CPU and several (5) php-fpm instances using 100% CPU. | |
| Jul 6 at 13:07 | comment | added | Phill Ash | I found this thread serverfault.com/questions/823121/… and followed it's guidance. I now have apache using php-fpm. Let's see if this helps. I'm avoiding changing it to the answer because other users have suggest it's merely kicking the can down the road whilst acknowledging it should be done anyway. Others suggest simply moving off apache onto nginx | |
| Jul 6 at 12:31 | comment | added | Phill Ash | I have a feeling I might have found what's really creating the problem with the database. 14886 0.37s 3.91s -3084K 0K 0K 0K -- - 51 S 3 170% mysqld 23900 0.27s 2.30s 2048K 1940K 0K 0K -- - 1 R 2 100% apache2 24106 0.38s 1.96s 0K 4K 0K 0K -- - 1 R 3 93% apache2 23893 0.12s 2.08s 0K 0K 0K 4K -- - 1 S 1 87% apache2 23988 0.30s 1.74s 0K 0K 0K 0K -- - 1 R 3 81% apache2 There are another 10 apache processes all running 81% CPU. | |
| Jul 6 at 12:25 | comment | added | Phill Ash | So I thought I had identified the issue. One of our wordpress backup tools updraftplus had been attempting to backup our site. It looked like the backup hadn't happened for several months and I discovered that the remote storage wasn't accessible. I renewed the fee payment for it (it was on an old payment card) backed the site up manually and the CPU seemed to drop to normal. I just had a look at our monitoring today though and I can see several high CPU alerts. It's possible that this is the backup happening since it's been reinstated... investigating. | |
| S Jul 5 at 23:18 | review | First questions | |||
| Jul 11 at 20:11 | |||||
| S Jul 5 at 23:18 | history | asked | Phill Ash | CC BY-SA 4.0 |