OK, can anyone estimate for me what size the swap log is likely to be if
there's 8GB of cached objects on that partitiion? Maybe 300K isn't
enough to write the swap log and this is causing the problem. Can anyone
see a way to get objective about this without replacing the RAM at this
stage?
Regards,
Alastair Waddell wrote:
>
> Hi, this appears to have been covered before with suggestions of memory
> failure but I'm half expecting something else to be behind it.
>
> I've just installed a couple of new scsi drives to complement the 9GB
> already installed. The IDE containing the O/S decided to shit itself at
> the end of my good work, and after a weekend of 'ide reset timeout' woes
> with a new IDE, I upgraded to linux 2.2.2 and all recommended utils.
> disabling DMA probes 'and stuff' got me mobile again but alas, with my
> new squid installation and config, the swap.state file on my 9GB scsi
> (somehow) got zeroed.
>
> Now I seem unable to rebuild my cache due to the violation below.
>
> You guessed it - I also installed another 128MB of SDRAM in amongst
> those scsi cables and I'm hanging out for any reason not to have to
> debug my RAM.
>
> I thought maybe it could be a malloc problem due to changes in /proc or
> something. Anything.
>
> There's only 300KB free on that cache partition (Max Swap size: 8294400
> KB) and I've no idea what size the final swap file would be (although it
> was obviously there before).
>
> Regards,
>
> 1999/03/02 05:30:25| Starting Squid Cache version 2.1.PATCH2 for
> i686-pc-linux-gnu...
> 1999/03/02 05:30:25| Process ID 4590
> 1999/03/02 05:30:25| With 1024 file descriptors available
> 1999/03/02 05:30:25| Performing DNS Tests...
> 1999/03/02 05:30:25| Successful DNS name lookup tests...
> 1999/03/02 05:30:25| helperOpenServers: Starting 5 'dnsserver' processes
> 1999/03/02 05:30:26| snmpInit: WARNING: communities not defined.
> 1999/03/02 05:30:26| Unlinkd pipe opened on FD 14
> 1999/03/02 05:30:26| Swap maxSize 8294400 KB, estimated 638030 objects
> 1999/03/02 05:30:26| Target number of buckets: 12760
> 1999/03/02 05:30:26| Using 16384 Store buckets, replacement runs every 5
> seconds
> 1999/03/02 05:30:26| Max Mem size: 8192 KB
> 1999/03/02 05:30:26| Max Swap size: 8294400 KB
> 1999/03/02 05:30:26| Rebuilding storage in Cache Dir #0 (DIRTY)
> 1999/03/02 05:30:26| Loaded Icons.
> 1999/03/02 05:30:26| Accepting HTTP connections on port 3128, FD 36.
> 1999/03/02 05:30:26| Accepting ICP messages on port 3130, FD 37.
> 1999/03/02 05:30:26| Accepting SNMP messages on port 3401, FD 38.
> 1999/03/02 05:30:26| Ready to serve requests.
> 1999/03/02 05:30:26| 65536 files opened so far.
> 1999/03/02 05:30:26| 131072 files opened so far.
> 1999/03/02 05:30:26| 196608 files opened so far.
> FATAL: Received Segment Violation...dying.
> 1999/03/02 05:30:26| Not currently OK to rewrite swap log.
> 1999/03/02 05:30:26| storeDirWriteCleanLogs: Operation aborted.
> CPU Usage: 91.080 seconds
> Maximum Resident Size: 0 KB
> Page faults with physical i/o: 474840
> Memory usage for squid via mallinfo():
> total space in arena: 20291 KB
> Ordinary blocks: 20276 KB 1 blks
> Small blocks: 0 KB 0 blks
> Holding blocks: 336 KB 2 blks
> Free Small blocks: 0 KB
> Free Ordinary blocks: 14 KB
> Total in use: 20612 KB 102%
> Total free: 14 KB 0%
>
> --
> ----------------------------------------------------------------------
> Alastair Waddell o Tel +61 3 96400400
> Legion Internet
> Queen Street, Melbourne o Full featured VISP Facility
>
> Virtual Services + DNS Maintenance + ISP Co-location + Internetworking
-- ---------------------------------------------------------------------- Alastair Waddell o Tel +61 3 96400400 Legion Internet Queen Street, Melbourne o Full featured VISP Facility Virtual Services + DNS Maintenance + ISP Co-location + InternetworkingReceived on Tue Mar 02 1999 - 04:52:26 MST
This archive was generated by hypermail pre-2.1.9 : Tue Dec 09 2003 - 16:45:06 MST