Hi all, We have a squid running for quite a few years and with the increase of traffic we noticed a bit of I/O hammering on the squid server (local disks). For some testing, I've made a small 1.2GB tmpfs and pointed cache_dir to it so that our cache would be in the 'ramdrive'. This did help a lot with I/O, but squid eventually once in a while implodes when it fills 1.2GB of the ramdrive (cache_dir is configured for 1G, leaving 200MB free). This is on a CentOS based system: [####EDIT#### squid]# rpm -qi squid Name : squid Relocations: (not relocatable) Version : 3.1.23 Vendor: CentOS Release : 9.el6 Build Date: Fri 24 Jul 2015 09:59:03 AM WEST Install Date: Tue 13 Oct 2015 02:31:20 PM WEST Build Host: c6b8.bsys.dev.centos.org Group : System Environment/Daemons Source RPM: squid-3.1.23 -9.el6.src.rpm Size : 6649558 License: GPLv2 and (LGPLv2+ and Public Domain) Signature : RSA/SHA1, Fri 24 Jul 2015 09:39:20 PM WEST, Key ID 0946fca2c105b9de Packager : CentOS BuildSystem <http://bugs.centos.org> URL : http://www.squid-cache.org Summary : The Squid proxy caching server Description : Anyone can help? Imagine I want to keep working on a ramdrive with 1.5G. How much can I get cache_dir configured with a safe margin so it doesnt fill the space? Any options I should look over again which might help me diagnose if it happens again? Kindest Regards, NMM -- Nelson Manuel Marques <nmm@xxxxxxxxxxx> Administrador de Sistemas Eurotux Informática S.A. | www.eurotux.com (t) +351 253 680 300
Attachment:
signature.asc
Description: This is a digitally signed message part
_______________________________________________ squid-users mailing list squid-users@xxxxxxxxxxxxxxxxxxxxx http://lists.squid-cache.org/listinfo/squid-users
