-1

My Google Cloud SQL instance stopped responding. I changed the size from a micro instance to standard instance and then the memory just shot to the ceiling and stayed there. The restart button is disabled and when trying to do mysql --host=___.___.__.___ --user=root --password it's unable to connect.

enter image description here

enter image description here

enter image description here

What is the recommended procedure to get my instance stable and accessible again? All sites connected simply times out trying to connect to that DB.

Update:

A few hours after the DB turned into a vegetable state, I was able to connect to it again, but memory usage is still extremely high and I'm still only able to delete it, can't click edit, imoprt, export, restart or change permissions. I can add permissions, but the save button is greyed out.

Here's a screen grab of the logs:

enter image description here

enter image description here

Running show processlist; only shows two processes, I've killed both processes and stopped all instances connected to it just to be sure it's nothing I'm running that's killing the DB.

Memory usage is still climbing as shown in the screenshot even though I've killed all processes:

enter image description here

Jan Vladimir Mostert
  • 553
  • 5
  • 19
  • 1
    What do the mysql logs say?Are you connecting to the correct IP address ? – user9517 Dec 28 '15 at 08:16
  • Yes, I've triple checked that I'm connecting to the right IP address and that my instance from which I'm connecting has access. I've since managed to get access to the DB and the websites are back up and running without changing anything, but still, the only option available on the instance is to delete it (restart, export, import and edit are still greyed out). How do I access the logs? – Jan Vladimir Mostert Dec 28 '15 at 08:24
  • Updated with screenshots of the logs and more information. – Jan Vladimir Mostert Dec 28 '15 at 10:00
  • 1
    The action controls being disabled might be a caching issue. Can you try a "Empty Cache & Reload" (http://www.thewindowsclub.com/empty-cache-hard-reload-chrome) and see if that helps? We have a fix rolling out after the holidays. – Vadim Dec 28 '15 at 17:47
  • 1
    Note that a fraction of the total memory is allocated to the InnoDB buffer pool so it's normal to see the memory usage jump when going to a higher tier with more memory. Did it keep going up after the jump? – Vadim Dec 28 '15 at 17:58
  • 1
    You can see the message "Initializing buffer pool" in the screenshot with the allocated memory. – Vadim Dec 28 '15 at 18:01
  • Yes, it kept on going up and up until it hit its ceiling. I deleted the instance in the meantime since I couldn't use it and migrated to a new instance which is running smooth. If I run into the same issue with controls being disabled, I'll try clearing the cache. Thanks Vadim! – Jan Vladimir Mostert Dec 28 '15 at 19:11
  • Let us [continue this discussion in chat](http://chat.stackexchange.com/rooms/33534/discussion-between-vadim-and-jan-vladimir-mostert). – Vadim Dec 28 '15 at 19:27
  • I'm voting to close this question as off-topic because it needs to be addressed to Google Support. – EEAA Jan 09 '16 at 22:13
  • If this is considered off-topic, where do you suggest I post it? There's no direct line for support unless you buy a support package which is overkill for supporting personal sites. – Jan Vladimir Mostert Jan 10 '16 at 04:18

1 Answers1

0

The UI bug that made it look like the size increase was still happening has been fixed by Google.

The not being able to connect to the instance was caused by a lot of bots hammering the site at the same time using up all the connections, but since the UI was in a disabled state, my assumption was that it was stuck in the upgrade process.

Jan Vladimir Mostert
  • 553
  • 5
  • 19