You can use the null cache_dir type to store the cache objects in memory,
but I understand some OSes have (had?) trouble dealing with 1gb+
processes. Linux also has tmpfs you can use to make the page cache a
giant in-memory filesystem.
You will need to set the cache_log to /dev/null and disable the other
logs. Also, set your filesystems to drop to read-only on errors.
-- Brian
On Wednesday 26 June 2002 09:28 am, you wrote:
> Hi!
> Is it possible to run squid only on memory for sometime ... so
> that in case of a hard disk failure the service is not disrupted during
> that period of time ... loss of data or performance degradation during
> that period is not a problem ...
>
> I had posted this message earlier but didnt get any replies ... so I
> thought I will repost it ...
Received on Wed Jun 26 2002 - 09:35:08 MDT
This archive was generated by hypermail pre-2.1.9 : Tue Dec 09 2003 - 17:08:49 MST