On 07.07.2014, at 20:17, Levente Uzonyi leves@elte.hu wrote:
The usual thing happened again: apache ran out of worker processes. The php running mantis can consume all 200 workers somehow. I logged in today and restarted php which resolved the problem. Since there was no mail about the downtime - even though it was about eight and a half hours long - I didn't bother writing a mail about this "well known" error. To prevent such downtimes again, I've set up a cron job which restarts the phpmantis service every hour.
On Mon, 7 Jul 2014, Tobias Pape wrote:
Dear all,
while investigating https://github.com/dalehenrich/metacello-work/issues/255#issuecomment-482031...
The build was started during the downtime, so that's why it failed. This is another reason why we should migrate the source.squeak.org image next. :)
Note that the other half of the symptoms was not being able to download the image from http://ftp.squeak.org, so migrating the file serving should also be high-priority.
I explored Box 2 and found that only
4 GB
AFAIK this box has always had such low empty disk space, because there are local backups of the local services (IIRC 5 snapshots each about 20 GB).
Best -Tobias