We are running squid in a vserver, but this does not work very reliably,
because the squid process crashes quite often. It seems that this happens
every time when squid is using more than 256 MB of cache memory - the
limit is set to a very high value in squid.conf, 24 GB.
Do vservers have a memory limit for processes so that a process that needs
more is killed? Or does somebody have an idea what other vserver features
can be responsible for the crashes?
We have tested the squid setup before on another Linux system, without
vserver, and the crashes did not occur there ...
Received on Mon Jul 1 19:40:14 2013