Search took 0.19 seconds

Search for '#mokum-outage' in texts

User avatar
» posted to mokum

Mokum had an outage that lasted for two hours. The root cause was that we had a massive misconfiguration that worked until you don't touch anything, and we did exactly that while trying to upgrade Ruby interpreter. Good news is that a lot of old cruft was cleaned up during firefighting, and Mokum is back online. Bad news is that it's not stable enough so tomorrow we need to find out what exactly is broken, and fix it. Sorry for the inconvenience. #mokum-outage

Comment

I see a very small number of occasional errors (two in past 30 minutes). They go away after reload.

 ‎· псы в рапиде
Comment

Thank you for all of your hard work, and letting us know :)

 ‎· Jennifer D. 11
Comment

Yes, thank you!

 ‎· Galadriel
Comment

Heads up: there is a chance that something will break again, but I can't really promise anything.

 ‎· псы в рапиде 4
Comment

Me either.

 ‎· Meg Vmeg
User avatar
» posted to mokum-support, mokum, and support

Public service announcement: tomorrow, Sep 28, Mokum website will be unavailable for some time between 3:00 AM and 5:00 AM UTC due to maintenance at our hosting provider. Expected downtime is no more than 40 minutes. Thank you for your attention.

Comment

details from Hetzner - https://www.hetzner-status.de/en.html#7859

 ‎· заёбы завтрашнего дня
Comment

(both Mokum servers will be affected, each could be down for up to 20 minutes, thus total downtime 40 minutes)

 ‎· заёбы завтрашнего дня
Comment

...and we had two more outages, yesterday during the day and this night. https://www.hetzner-status.de/ #mokum-outage

 ‎· псы в рапиде
User avatar
» posted to support

Today between 02:35 UTC and 05:05 UTC Mokum was down. Sorry about that! The reason behind that is one of the hardest problems in computer science, namely shortage of disk space due to forgotten log file. @ayoshi and @haron handled this problem, and service was restored (thank you, guys!) We'll be improving disk space situation by moving half of images to the second server, but it requires some changes to Mokum code base). 2.5 hours of downtime gives us ~99.97% of yearly uptime so far. Thank you for your support, #mokum-outage

Comment

this needs to be a horror anime: The Log File That Ate Neo-Tokyo!

 ‎· LibSkrat 8
Comment

@libskrat: https://www.youtube.com/watch?v=Xt4ua_ZNoD0

 ‎· псы в рапиде 5

1 2 3 4 5 6 7 8 9 10