disk error on squid

proxy squid i used is always restart. then stop. when i look in log i get this error.
Code:
storeDiskdSend OPEN: (35) Resource temporarily unavailable
storeDiskdSend: msgsnd: (35) Resource temporarily unavailable
assertion failed: diskd/store_io_diskd.c:565: "++send_errors < 100"
i use 250gig hdd i split the for caching 10 gig each on squid i use setting
Code:
cache_dir diskd /cache0 6900 18 256 Q1=72 Q2=64
cache_dir diskd /cache1 6900 18 256 Q1=72 Q2=64
cache_dir diskd /cache2 6900 18 256 Q1=72 Q2=64
cache_dir diskd /cache3 6900 18 256 Q1=72 Q2=64
cache_dir diskd /cache4 6900 18 256 Q1=72 Q2=64
cache_dir diskd /cache5 6900 18 256 Q1=72 Q2=64
cache_dir diskd /cache6 6900 18 256 Q1=72 Q2=64
cache_dir diskd /cache7 6900 18 256 Q1=72 Q2=64
cache_dir diskd /cache8 6900 18 256 Q1=72 Q2=64
cache_dir diskd /cache9 6900 18 256 Q1=72 Q2=64
cache_dir diskd /cache10 6900 18 256 Q1=72 Q2=64
cache_dir diskd /cache11 6900 18 256 Q1=72 Q2=64
cache_dir diskd /cache12 6900 18 256 Q1=72 Q2=64
ram 1 gig ddr1.
help me please
 
i try to recompile kernel and squid again and i get error
Code:
2009/06/14 12:16:12| Starting Squid Cache version 2.7.STABLE6 for i386-unknown-freebsd7.0...
2009/06/14 12:16:12| Process ID 16658
2009/06/14 12:16:12| With 11095 file descriptors available
2009/06/14 12:16:12| Using kqueue for the IO loop
2009/06/14 12:16:12| DNS Socket created at 0.0.0.0, port 49764, FD 6
2009/06/14 12:16:12| Adding nameserver 192.168.100.101 from /etc/resolv.conf
2009/06/14 12:16:12| logfileOpen: opening log /var/log/squid/access.log
2009/06/14 12:16:12| Unlinkd pipe opened on FD 11
2009/06/14 12:16:12| Swap maxSize 91852800 + 786432 KB, estimated 0 objects
2009/06/14 12:16:12| Target number of buckets: 356304
2009/06/14 12:16:12| Using 524288 Store buckets
2009/06/14 12:16:12| Max Mem  size: 786432 KB
2009/06/14 12:16:12| Max Swap size: 91852800 KB
2009/06/14 12:16:12| Store logging disabled
2009/06/14 12:16:12| Rebuilding storage in /cache0 (DIRTY)
2009/06/14 12:16:12| Rebuilding storage in /cache1 (DIRTY)
2009/06/14 12:16:12| Rebuilding storage in /cache2 (DIRTY)
2009/06/14 12:16:12| Rebuilding storage in /cache3 (DIRTY)
2009/06/14 12:16:12| Rebuilding storage in /cache4 (DIRTY)
2009/06/14 12:16:12| Rebuilding storage in /cache5 (DIRTY)
2009/06/14 12:16:12| Rebuilding storage in /cache6 (DIRTY)
2009/06/14 12:16:12| Rebuilding storage in /cache7 (DIRTY)
2009/06/14 12:16:12| Rebuilding storage in /cache8 (DIRTY)
2009/06/14 12:16:12| Rebuilding storage in /cache9 (DIRTY)
2009/06/14 12:16:12| Rebuilding storage in /cache10 (DIRTY)
2009/06/14 12:16:12| storeDiskdInit: msgget: (28) No space left on device
FATAL: msgget failed
Squid Cache (Version 2.7.STABLE6): Terminated abnormally.
CPU Usage: 0.031 seconds = 0.008 user + 0.023 sys
Maximum Resident Size: 5280 KB
Page faults with physical i/o: 0
 
2009/06/14 12:16:12| storeDiskdInit: msgget: (28) No space left on device

You are running out of disk space. Remove unwanted files especially purge squid cache by deleting files or using squid client. df and du command will provide info about disk space usage and estimate file space usage respectively.
 
i think the problem is not because i need to purge the data. when i use "df -h" coomand it show

Code:
/dev/ad6s2d    9.5G    546M    8.2G     6%    /cache0
/dev/ad6s2e    9.5G    572M    8.1G     6%    /cache1
/dev/ad6s3g    9.5G    593M    8.1G     7%    /cache10
/dev/ad6s3h    9.5G    616M    8.1G     7%    /cache11
/dev/ad6s3a    9.5G    537M    8.2G     6%    /cache12
/dev/ad6s2f    9.5G    554M    8.2G     6%    /cache2
/dev/ad6s2g    9.5G    535M    8.2G     6%    /cache3
/dev/ad6s2h    9.5G    574M    8.1G     6%    /cache4
/dev/ad6s2a    9.5G    532M    8.2G     6%    /cache5
/dev/ad6s2b    9.5G    566M    8.1G     6%    /cache6
/dev/ad6s3d    9.5G    578M    8.1G     6%    /cache7
/dev/ad6s3e    9.5G    569M    8.1G     6%    /cache8
/dev/ad6s3f    9.5G    627M    8.1G     7%    /cache9
 
How big are the swap.state files? Squid will try to regenerate them when reinitialising 'dirty' cache directories, which means there will be a second swap.state file being built in place, which can be as big as the original. If there's no room for that second swap.state file, cleaning will fail and squid will terminate.

Any other partition full, like /tmp or /var/log?
 
Code:
/dev/ad6s1a    496M    209M    247M    46%    /
devfs          1.0K    1.0K      0B   100%    /dev
/dev/ad6s3b    9.5G     36K    8.7G     0%    /cache
/dev/ad6s2d    9.5G    546M    8.2G     6%    /cache0
/dev/ad6s2e    9.5G    572M    8.1G     6%    /cache1
/dev/ad6s3g    9.5G    593M    8.1G     7%    /cache10
/dev/ad6s3h    9.5G    616M    8.1G     7%    /cache11
/dev/ad6s3a    9.5G    537M    8.2G     6%    /cache12
/dev/ad6s2f    9.5G    554M    8.2G     6%    /cache2
/dev/ad6s2g    9.5G    535M    8.2G     6%    /cache3
/dev/ad6s2h    9.5G    574M    8.1G     6%    /cache4
/dev/ad6s2a    9.5G    532M    8.2G     6%    /cache5
/dev/ad6s2b    9.5G    566M    8.1G     6%    /cache6
/dev/ad6s3d    9.5G    578M    8.1G     6%    /cache7
/dev/ad6s3e    9.5G    569M    8.1G     6%    /cache8
/dev/ad6s3f    9.5G    627M    8.1G     7%    /cache9
/dev/ad6s4d     47G    4.0K     44G     0%    /data
/dev/ad6s4e     40G     71M     36G     0%    /ext
/dev/ad6s1e    3.3G    1.5G    1.5G    50%    /usr
/dev/ad6s1d    496M     66M    390M    14%    /var

i think /var/log almost full. i allready check on swap state its only 6.9mb. is there is a way to make it clean.
is there is way to uninstall squids using ports using "make fetch package install".
 
neo_leopard said:
Code:
i think /var/log almost full. i allready check on swap state its only 6.9mb. is there is a way to make it clean.
is there is way to uninstall squids using ports using "[B]make fetch package install[/B]".[/QUOTE]


Yes, quick solution delete files and dir from /var/log:
[CODE]cd /var/log
rm unwanted.log
rm unwanted.log.1
rm unwanted.log.1.gz

You must use newsyslog(8) program to keep tab on log file size. Also if squid cache is not important just delete all those swap dir and recreate dirs using squid -z option as root (this can be done with squid rc.d script itself also but I'm not sure).
 
it still show error but its different.

Code:
2009/06/15 16:18:12| storeDiskdSend: msgsnd: (35) Resource temporarily unavailable
2009/06/15 16:18:12| storeDiskdSend WRITE: (35) Resource temporarily unavailable
2009/06/15 16:18:12| storeSwapOutFileClosed: dirno 4, swapfile 00000061, errflag=-1
        (35) Resource temporarily unavailable
2009/06/15 16:18:12| storeDiskdSend: msgsnd: (35) Resource temporarily unavailable
2009/06/15 16:18:12| storeDiskdSend UNLINK: (35) Resource temporarily unavailable
2009/06/15 16:18:12| storeDiskdSend: msgsnd: (35) Resource temporarily unavailable
2009/06/15 16:18:12| storeDiskdSend WRITE: (35) Resource temporarily unavailable
2009/06/15 16:18:12| storeSwapOutFileClosed: dirno 6, swapfile 00000061, errflag=-1
        (35) Resource temporarily unavailable

i allready delete "/var/log" and create the squid folder.
is there is away make the partition clean.
 
Did you read what I posted earlier? Stop squid and go to each /cache?? dir and delete everything using rm command. Run above squid command.
 
Maybe the problem is not disk space, but diskd itself. To use diskd you need kernel modifications (you may get away with changing stuff in sysctl.conf/loader.conf nowadays). Try changing diskd to ufs or aufs and see if that works.
 
Back
Top