|
From: Dave <gh...@gm...> - 2006-09-05 13:26:33
|
Hey folks, I'm running a 1.2.10 as a firewall, with squid & Squidguard, and I've run into a slight prob... How can I disable the on-disk cache for squid? It seems to be only an option at the start when building from source to disable it. Is there some trick I'm missing perhaps? Thoughts? Thanks in advance, and keep up the good work, DL rocks! :) Dave Mullen |
|
From: Bruce S. <bw...@ar...> - 2006-09-05 14:38:58
|
> I'm running a 1.2.10 as a firewall, with squid & Squidguard, and I've > run into a slight prob... How can I disable the on-disk cache for > squid? It seems to be only an option at the start when building from > source to disable it. Is there some trick I'm missing perhaps? > Thoughts? I'm running squid on one DL box without any hard drive. I just set the cache really small in the squid.conf: cache_dir ufs /var/squid/cache 16 4 16 And I also turned off the log files: cache_access_log none cache_store_log none - BS |
|
From: Dave <gh...@gm...> - 2006-09-05 17:15:17
|
Hmmm. Might I ask what you did for the cache_swap_log ? That's one file I've not got contained either! It just seems to grow unless I do a squid -k rotate. /sigh Dave On 9/5/06, Bruce Smith <bw...@ar...> wrote: > > > I'm running a 1.2.10 as a firewall, with squid & Squidguard, and I've > > run into a slight prob... How can I disable the on-disk cache for > > squid? It seems to be only an option at the start when building from > > source to disable it. Is there some trick I'm missing perhaps? > > Thoughts? > > I'm running squid on one DL box without any hard drive. > I just set the cache really small in the squid.conf: > > cache_dir ufs /var/squid/cache 16 4 16 > > And I also turned off the log files: > > cache_access_log none > cache_store_log none > > - BS > > > > ------------------------------------------------------------------------- > Using Tomcat but need to do more? Need to support web services, security? > Get stuff done quickly with pre-integrated technology to make your job > easier > Download IBM WebSphere Application Server v.1.0.1 based on Apache Geronimo > http://sel.as-us.falkag.net/sel?cmd=lnk&kid=120709&bid=263057&dat=121642 > _______________________________________________ > Devil-linux-discuss mailing list > Dev...@li... > https://lists.sourceforge.net/lists/listinfo/devil-linux-discuss > |
|
From: Bruce S. <bw...@ar...> - 2006-09-05 17:28:17
|
> Hmmm. Might I ask what you did for the cache_swap_log ? That's one > file I've not got contained either! I didn't do anything with it, just left the defaults in squid.conf. Maybe it doesn't get very large if you set the cache_dir small, since it's "used to rebuild the cache during startup". If there isn't much cache, it seems like there wouldn't be much metadata in the swap log. My entire /var/squid/cache directory is 22MB. > It just seems to grow unless I do a squid -k rotate. /sigh That would be easy to script in a crontab, if you had to. - BS |
|
From: Dave <gh...@gm...> - 2006-09-05 17:56:10
|
Yeah, with 500 people going through it however, that cache_swap_log continues to grow, just under 1k/sec. Afterhours it doesn't gain much at all. So far I've not seen it stop, just gain space. The squid -k rotate is what I used to do, but now I'm using squidGuard with a huge blacklist service ( http://urlblacklist.com/ ) so with 5 child processes it takes about a minute of downtime for it to be ready to roll again. For right now, this hasn't been a big issue, as it's an HA cluster, and I can just rebuild it on the secondary node, and failover to it. Dave On 9/5/06, Bruce Smith <bw...@ar...> wrote: > > > Hmmm. Might I ask what you did for the cache_swap_log ? That's one > > file I've not got contained either! > > I didn't do anything with it, just left the defaults in squid.conf. > > Maybe it doesn't get very large if you set the cache_dir small, since > it's "used to rebuild the cache during startup". If there isn't much > cache, it seems like there wouldn't be much metadata in the swap log. > > My entire /var/squid/cache directory is 22MB. > > > It just seems to grow unless I do a squid -k rotate. /sigh > > That would be easy to script in a crontab, if you had to. > > - BS > > > > ------------------------------------------------------------------------- > Using Tomcat but need to do more? Need to support web services, security? > Get stuff done quickly with pre-integrated technology to make your job > easier > Download IBM WebSphere Application Server v.1.0.1 based on Apache Geronimo > http://sel.as-us.falkag.net/sel?cmd=lnk&kid=120709&bid=263057&dat=121642 > _______________________________________________ > Devil-linux-discuss mailing list > Dev...@li... > https://lists.sourceforge.net/lists/listinfo/devil-linux-discuss > |
|
From: Chad M. <ch...@th...> - 2006-09-06 01:48:49
|
Dave wrote: > Yeah, with 500 people going through it however, that cache_swap_log > continues to grow, just under 1k/sec. Afterhours it doesn't gain much > at all. So far I've not seen it stop, just gain space. The squid -k > rotate is what I used to do, but now I'm using squidGuard with a huge > blacklist service ( http://urlblacklist.com/ ) so with 5 child processes > it takes about a minute of downtime for it to be ready to roll again. > For right now, this hasn't been a big issue, as it's an HA cluster, and > I can just rebuild it on the secondary node, and failover to it. How about: ln -s /dev/null cache_swap_log Chad Martin |
|
From: Bruce S. <bw...@ar...> - 2006-09-05 18:06:31
|
Yeah, I don't have many users on my diskless squid box. It seems like with 500 users, big fast hard drive(s) to hold a lot of cache would be worth the investment in bandwidth savings. - BS > Yeah, with 500 people going through it however, that cache_swap_log > continues to grow, just under 1k/sec. Afterhours it doesn't gain much > at all. So far I've not seen it stop, just gain space. The squid -k > rotate is what I used to do, but now I'm using squidGuard with a huge > blacklist service ( http://urlblacklist.com/ ) so with 5 child > processes it takes about a minute of downtime for it to be ready to > roll again. For right now, this hasn't been a big issue, as it's an > HA cluster, and I can just rebuild it on the secondary node, and > failover to it. > > Dave > > On 9/5/06, Bruce Smith <bw...@ar...> wrote: > > Hmmm. Might I ask what you did for the > cache_swap_log ? That's one > > file I've not got contained either! > > I didn't do anything with it, just left the defaults in > squid.conf. > > Maybe it doesn't get very large if you set the cache_dir > small, since > it's "used to rebuild the cache during startup". If there > isn't much > cache, it seems like there wouldn't be much metadata in the > swap log. > > My entire /var/squid/cache directory is 22MB. > > > It just seems to grow unless I do a squid -k rotate. /sigh > > That would be easy to script in a crontab, if you had to. |
|
From: Dave <gh...@gm...> - 2006-09-05 18:23:13
|
Yeah, I mentioned that. Several times actually! /sigh I've been looking to replace these boxes with a box with a raid array of some sort, just for it's cache, something we don't care about at all. Security here would rather not let it have any writable media though. Thanks again for your help though! Dave On 9/5/06, Bruce Smith <bw...@ar...> wrote: > > Yeah, I don't have many users on my diskless squid box. > > It seems like with 500 users, big fast hard drive(s) to hold a lot of > cache would be worth the investment in bandwidth savings. > > - BS > > > > Yeah, with 500 people going through it however, that cache_swap_log > > continues to grow, just under 1k/sec. Afterhours it doesn't gain much > > at all. So far I've not seen it stop, just gain space. The squid -k > > rotate is what I used to do, but now I'm using squidGuard with a huge > > blacklist service ( http://urlblacklist.com/ ) so with 5 child > > processes it takes about a minute of downtime for it to be ready to > > roll again. For right now, this hasn't been a big issue, as it's an > > HA cluster, and I can just rebuild it on the secondary node, and > > failover to it. > > > > Dave > > > > On 9/5/06, Bruce Smith <bw...@ar...> wrote: > > > Hmmm. Might I ask what you did for the > > cache_swap_log ? That's one > > > file I've not got contained either! > > > > I didn't do anything with it, just left the defaults in > > squid.conf. > > > > Maybe it doesn't get very large if you set the cache_dir > > small, since > > it's "used to rebuild the cache during startup". If there > > isn't much > > cache, it seems like there wouldn't be much metadata in the > > swap log. > > > > My entire /var/squid/cache directory is 22MB. > > > > > It just seems to grow unless I do a squid -k rotate. /sigh > > > > That would be easy to script in a crontab, if you had to. > > > > > ------------------------------------------------------------------------- > Using Tomcat but need to do more? Need to support web services, security? > Get stuff done quickly with pre-integrated technology to make your job > easier > Download IBM WebSphere Application Server v.1.0.1 based on Apache Geronimo > http://sel.as-us.falkag.net/sel?cmd=lnk&kid=120709&bid=263057&dat=121642 > _______________________________________________ > Devil-linux-discuss mailing list > Dev...@li... > https://lists.sourceforge.net/lists/listinfo/devil-linux-discuss > |