Heya gert, df -h shows Filesystem Size Used Avail Use% Mounted on /dev/sda5 22G 6.3G 15G 31% / /dev/sda1 99M 36M 58M 39% /boot none 501M 0 501M 0% /dev/shm /dev/sda2 9.7G 2.9G 6.3G 32% /var/log/squid /dev/sdb1 68G 65M 64G 1% /var/squid/cache <-- cache dir I can ping/traceroute/nslookup/host everywhere Ive done that, ive even set it back to defaults on /usr/local/squid/cache (still says its dirty) Done all of that im afraid... Ross -----Original Message----- From: Gert Brits [mailto:gbrits@xxxxxxxxxxxxxxxxxx] Sent: 25 July 2005 14:43 To: squid-users@xxxxxxxxxxxxxxx Subject: RE: Problem with squid... Dirty Cache constant restart loop.... Hi There 1.) Check your disk space ( your cache partition if it is on its own ). 2.) Check if you have a proper DNS server that can resolve DNS names under /etc/resolve.conf 3.) Remove your old cache under your cache directory. 4.) squid -z 5.) restart squid. -----Original Message----- From: Ross McInnes [mailto:sysrm@xxxxxxxxxxxxxxx] Sent: 25 July 2005 03:27 PM To: squid-users@xxxxxxxxxxxxxxx Subject: Problem with squid... Dirty Cache constant restart loop.... Hi all. Been running squid for years, never really had any problems till today. Everything seems to have been running fine till just before 12pm.. Then someone comes in and asks if the net is down.... Sure enough squid has died and is no longer running on the server. I do a service squid start Try to connect to the internet again... Nothing. No username/password request etc. Tail -f /var/log/squid/cache.log I get... 2005/07/25 15:10:14| Starting Squid Cache version 2.4.STABLE6 for i686-pc-linux-gnu... 2005/07/25 15:10:14| Process ID 5379 2005/07/25 15:10:14| With 1024 file descriptors available 2005/07/25 15:10:14| Performing DNS Tests... 2005/07/25 15:10:14| Successful DNS name lookup tests... 2005/07/25 15:10:14| DNS Socket created on FD 4 2005/07/25 15:10:14| Adding nameserver 127.0.0.1 from /etc/resolv.conf 2005/07/25 15:10:14| helperOpenServers: Starting 20 'squirm' processes 2005/07/25 15:10:14| helperOpenServers: Starting 15 'ncsa_auth' processes 2005/07/25 15:10:17| Unlinkd pipe opened on FD 44 2005/07/25 15:10:17| Swap maxSize 61440000 KB, estimated 4726153 objects 2005/07/25 15:10:17| Target number of buckets: 236307 2005/07/25 15:10:17| Using 262144 Store buckets 2005/07/25 15:10:17| Max Mem size: 8192 KB 2005/07/25 15:10:17| Max Swap size: 61440000 KB 2005/07/25 15:10:17| Rebuilding storage in /var/squid/cache (DIRTY) 2005/07/25 15:10:17| Using Least Load store dir selection 2005/07/25 15:10:17| Set Current Directory to /var/squid/cache Repeated over and over and over etc. Ok, I figure that the cached files are not happy, So I remove the cache files, then set them back up using -z still the same problem. Rebooted the box, removed squirm and ncsa_auth, still the same... Im looking to remove the cache next to see if that helps it. Looked on google etc for this problem and no one seems to have an answer.... Any ideas? Many Thanks Ross
