> On Thursday 13 October 2005 13:52, Rodrigo A B Freire wrote:
> > Not to mention that when a squid -k rotate is issued (or a -k
> > reconfigure), the process may grow up to 2x the amount of its size
> > for a few seconds... If you don't have swap enough to buffer this
> > grow, the proxy will die miserabily out of memory with a FATAL:
> > xmalloc: Unable to allocate xxx bytes.
I don't see this that many problematic, especially on linux, which is very
effective in these case, so the real memory usage does not grow 2 times.
On 13.10 14:01, Raymond A. Meijer wrote:
> Good point...but I think I saw it recommended somewhere? Henrik?
>
> Something about disabling swap and making sure Squid fits into the
> available memory...
it was probably the second part. there should be enough of memory for squid
to fit it all in. however, in some cases there MAY be other data stored on
swap. So, if you really have enough of memory (and you should have), you can
disable swap.
But this was not the case.
-- Matus UHLAR - fantomas, uhlar@fantomas.sk ; http://www.fantomas.sk/ Warning: I wish NOT to receive e-mail advertising to this address. Varovanie: na tuto adresu chcem NEDOSTAVAT akukolvek reklamnu postu. "The box said 'Requires Windows 95 or better', so I bought a Macintosh".Received on Fri Oct 14 2005 - 01:17:03 MDT
This archive was generated by hypermail pre-2.1.9 : Tue Nov 01 2005 - 12:00:04 MST