Re: [squid-users] Squid dies: no space left on device

You might take a look at your "swap.state" file (in your cache dir.) It
has been my experience that upon startup Squid makes a new copy of this
file. If it is fairly sizable, it will fail to come up if it can not
make room for the 2 copies of this file.

Peter Smith
Linux Systems Administrator
University of Texas Southwestern Medical Center at Dallas
(USA) 214 648 3111
peter.smith@utsouthwestern.edu

Winter Christian wrote:

>Hello everyone,
>
>I'm running Squid 2.2stable5 on Debian GNU Linux.
>Recently all my disk cache has filled up to 100%,
>although I have set a smaller Limit in squid.conf,
>and now squid dies a few seconds after being started.
>
>/etc/squid.conf has the entry
>cache_dir /var/spool/squid2 14000 64 256
>
>which should tell squid to expire old content if the used space
>reaches 14 GB, or am I wrong here?
>
>/proc/sys/fs/file-max is set to 16384
>
>df shows the following output:
>/dev/sdb1 17654736 16757928 0 100% /data/squid2
>
>so it is over the Limit of 14336000 Bytes set in squid.conf.
>
>How can I force squid to expire old content?
>
>I would be thankful for any advice, also any hint where to look
>for an answer.
>
>Thanx in advance
>Christian Winter
>
>
>
>--------------------------------------Output of
>cache.log---------------------------------------------
>2002/01/24 10:03:49| Starting Squid Cache version 2.2.STABLE5 for
>i386-debian-linux-gnu...
>2002/01/24 10:03:49| Process ID 31189
>2002/01/24 10:03:49| With 1024 file descriptors available
>2002/01/24 10:03:49| helperOpenServers: Starting 32 'dnsserver' processes
>2002/01/24 10:03:49| Swap maxSize 18944000 KB, estimated 1457230 objects
>2002/01/24 10:03:49| Target number of buckets: 29144
>2002/01/24 10:03:49| Using 32768 Store buckets, replacement runs every 2
>seconds
>2002/01/24 10:03:49| Max Mem size: 8192 KB
>2002/01/24 10:03:49| Max Swap size: 18944000 KB
>2002/01/24 10:03:49| Local cache digest enabled; rebuild/rewrite every
>3600/3600 sec
>2002/01/24 10:03:49| Rebuilding storage in Cache Dir #0 (DIRTY)
>2002/01/24 10:03:49| Rebuilding storage in Cache Dir #1 (DIRTY)
>2002/01/24 10:03:49| Set Current Directory to /var/spool/squid
>2002/01/24 10:03:49| Loaded Icons.
>2002/01/24 10:03:50| Accepting HTTP connections on port 3128, FD 63.
>2002/01/24 10:03:50| Accepting ICP messages on port 3130, FD 64.
>2002/01/24 10:03:50| Pinger socket opened on FD 66
>2002/01/24 10:03:50| NETDB state reloaded; 25 entries, 14 msec
>2002/01/24 10:03:50| Ready to serve requests.
>2002/01/24 10:03:51| 65536 Entries read so far.
>2002/01/24 10:03:52| 131072 Entries read so far.
>2002/01/24 10:03:52| 196608 Entries read so far.
>2002/01/24 10:03:53| 262144 Entries read so far.
>2002/01/24 10:03:54| 327680 Entries read so far.
>2002/01/24 10:03:55| 393216 Entries read so far.
>2002/01/24 10:03:56| 458752 Entries read so far.
>2002/01/24 10:03:57| icmpRecv: recv: (111) Connection refused
>2002/01/24 10:03:57| Closing Pinger socket on FD 66
>2002/01/24 10:03:57| diskHandleWrite: FD 39: disk write error: (28) No space
>left on device
>FATAL: Write failure -- check your disk space and cache.log
>Squid Cache (Version 2.2.STABLE5): Terminated abnormally.
>CPU Usage: 6.970 seconds
>Maximum Resident Size: 0 KB
>Page faults with physical i/o: 319
>Memory usage for squid via mallinfo():
> total space in arena: 47695 KB
> Ordinary blocks: 47519 KB 9985 blks
> Small blocks: 0 KB 0 blks
> Holding blocks: 1472 KB 4 blks
> Free Small blocks: 0 KB
> Free Ordinary blocks: 176 KB
> Total in use: 48991 KB 103%
> Total free: 176 KB 0%
>
Received on Thu Jan 24 2002 - 10:16:05 MST