AW: HELP: squid dies with "Unable to allocate 4096 blocks of 1 by te", I'm sorry

hello squid-users,

i have to apologize for not reading the faq 8.7! i'm so sorry!

faq 8.7 solves the problem. ulimit -aH shows me a maximum data-size of
131072. this is round about the size of the squid process before the
failure.
thanks all for help

martin

> -----Ursprüngliche Nachricht-----
> Von: Grabmann Martin [SMTP:martin.grabmann@rbg5.siemens.de]
> Gesendet am: Donnerstag, 30. März 2000 17:41
> An: 'squid-users@ircache.net'
> Betreff: HELP: squid dies with "Unable to allocate 4096 blocks of 1
> byte"
>
> hello squid-users,
>
> my squid runs well since a week, but today he suddenly died and was
> restarted by his father-process.
>
> my box is a HP-UX C110 with 256 of RAM. nothing else is running on this
> machine. the squid process was round about 130 MB of size as it died.
> i using squid 2.2.5, comiled with gcc 2.8.1 on HP-UX 10.20
>
> in the cache.log the following messages occurs
>
> 2000/03/30 14:27:52| sslReadClient: FD 26: read failure: (232) Connection
> reset by peer
> 2000/03/30 15:20:25| storeDirClean: /squid03/cache/03/D5: (12) Not enough
> space
> 2000/03/30 15:20:40| storeDirClean: /squid01/cache/04/D5: (12) Not enough
> space
> FATAL: xcalloc: Unable to allocate 4096 blocks of 1 bytes!
>
> Squid Cache (Version 2.2.STABLE5): Terminated abnormally.
> CPU Usage: 17081.840 seconds
> Maximum Resident Size: 128072 KB
> Page faults with physical i/o: 6
> Memory usage for squid via mallinfo():
> total space in arena: 129937 KB
> Ordinary blocks: 37523 KB 14183 blks
> Small blocks: 90800 KB 1770500 blks
> Holding blocks: 277 KB 17705 blks
> Free Small blocks: 338 KB
> Free Ordinary blocks: 999 KB
> Total in use: 128600 KB 99%
> Total free: 1337 KB 1%
> 2000/03/30 15:20:46| storeDirWriteCleanLogs: Starting...
> FATAL: xcalloc: Unable to allocate 16384 blocks of 1 bytes!
>
> my cache_dirs are configured like this
>
> cache_dir /squid01/cache 1000 32 256
> cache_dir /squid02/cache 1000 32 256
> cache_dir /squid03/cache 1000 32 256
> cache_mem 32 MB
>
> output of bdf -i
>
> # bdf -i
> Filesystem kbytes used avail %used iused ifree %iuse
> Mounted
> on
> /dev/vg00/lvol3 1363968 996109 345161 74% 41070 91962 31% /
> /dev/vg00/lvol1 99669 19359 70343 22% 21 16107 0%
> /stand
> /dev/vg00/lvol4 102400 9931 86739 10% 604 23116 3% /tmp
> /dev/dsk/c0t4d0 1956086 375408 1481946 20% 47 395169 0%
> /squid04
> /dev/dsk/c0t2d0 1956086 952505 940887 50% 165770 250894 40%
> /squid03
> /dev/dsk/c0t3d0 4194157 977834 3015354 24% 101172 804080 11%
> /squid02
> /dev/dsk/c0t1d0 4194157 1124394 2877934 28% 99461 767439 11%
> /squid01
>
> i think there should be enough space for squid in the memory and on the
> disks.
>
> questions:
>
> why squid dies?
> why only the directories /squid01 and squid03 are affected and not the
> /squid02 dir ?
> why the /squid01-directory is 1100 MB of size, although i gave it only
> 1000
> MB?
>
>
> thanks for help
> martin
Received on Thu Mar 30 2000 - 23:43:16 MST