Re: [squid-users] squid -k rotate restarts url_rewriters!

From: Adrian Chadd <[email protected]>
Date: Fri, 1 Feb 2008 11:04:22 +0900

On Thu, Jan 31, 2008, Chris Woodfield wrote:
> I just put a squid system with url_rewriter children into production.
> Alongside this we have a script that regularly runs squid -k rotate,
> then FTPs the log.1 files to a remote site for backup/processing.
>
> The issue I've noticed is that every time squid -k rotate is run,
> squid also stops and restarts all of its url_rewriter children. While
> this happens fairly quickly, it's a step I'd rather not have it do if
> necessary, particularly under the load these boxes are expected to be
> able to handle.
>
> Consider this a feature request (and if 3.0 does this already, please
> let me know): do not restart url_rewriter children on a USR1 signal
> (or whichever signal squid -k rotate sends in 2.6 STABLE18).

Well, I was thinking about splitting out some of the actions into
internal URLs rather than signals (as we're going to run out of
portable signals at some point) for various actions, including:

* rotate the normal logs, but not the swap logs
* rotate the swap logs
* gracefully restart the helpers
* re-load the ACL text files, but don't re-read the entire configuration

Adding these into Squid-2 and Squid-3 shouldn't be all that hard
(as practically the same calls are going to be made..)

Send me a private email if you'd like to discuss getting it done.

Adrian

-- 
- Xenion - http://www.xenion.com.au/ - VPS Hosting - Commercial Squid Support -
- $25/pm entry-level VPSes w/ capped bandwidth charges available in WA -
Received on Thu Jan 31 2008 - 18:53:20 MST

This archive was generated by hypermail pre-2.1.9 : Fri Feb 01 2008 - 12:00:05 MST