On Tue, 21 Apr 2009 14:58:22 +1200 (NZST), "Amos Jeffries" <squid3@xxxxxxxxxxxxx> wrote: [cut] > As a side issue: know who the maintainer is for slackware? I'm trying to > get in touch with them all. Sorry, here Squid was build from sources. The distro maintainer and more info (does not provide a binary Squid package) can be found here: http://bluewhite64.com (I'm still waiting for an official 64 bits Slackware port) >>> does deleting the swap.state file(s) when squid is stopped fix things? Apparently yes: /dev/sdb1 276G 225G 37G 87% /var/cache/proxy/cache1 /dev/sdc1 276G 225G 53G 87% /var/cache/proxy/cache2 /dev/sdd1 276G 225G 37G 87% /var/cache/proxy/cache3 /dev/sde1 276G 225G 37G 87% /var/cache/proxy/cache4 It's running OK again. Now, another strange log: 2009/04/21 00:26:25| commonUfsDirRebuildFromDirectory: Swap data buffer length is not sane. Should I decrease cache_dir sizes? >> The stranger think was store rebuild reporting > 100%. > > Yes, we have seen a similar thing long ago in testing. I'm trying to > remember and research what came of those. At present I'm thinking maybe it > had something to do with 32-bit/64-bit changes in distro build vs what the > cache was built with. Similar logs found here about memory usage (via mallinfo): Total in use: 1845425 KB 173% and sometimes negative values: total space in arena: -1922544 KB Ordinary blocks: -1922682 KB 49 blks Total in use: -1139886 KB 59% -- Herbert