Re: [squid-users] problems upgrading from squid 2.4stable6 to 2.5stabel2

> marc.elsen@imec.be wrote:
>
> > markus.rietzler@rzf.fin-nrw.de wrote:
> >
> > we have trouble with upgrading our squid to the most recent
> 2.5stabel2
> > version.
> >
> > after setting up and starting squid 2.5stabel2 there will
> be >90% CPU
> > load in top.
> > surfing will be very slow up to nearly not possible.
> >
> > we are now searching for a reason. squid 2.4stable6 is
> running without
> > any problems,
> > the 2.5 version on the same hardware can't be used.
> >
> > where using squid on hp-ux 11
>
>
> Anything special in cache.log ?
>
> How big is the cache ?
>
> How much physical ram do you have ?
>
> What is the output of ?
>
> # swapinfo -m
>
> What type of store access is being used (ufs,aufs,...) ?
>

B.11.00 A 9000/803
ram is 512 MB

where using aufs, cache_mem set to 32MB
cache_dir aufs /rzf/db/www/cache 512 16 256

where using a large number of acls.
the strange thing is, that 2.4 is working without problems, 2.5 have
such huge problems. nearly useless.

when using squid2.5 we had a lot of "(233) No buffer space available"
message. ok, they also appear with 2.4, but not that much.

there are at about 500 users setup for the cache. we are serving at
about 1 mio hits/day with traffic of 6 GB/day.

swapinfo -m says (for squid 2.4)
swapinfo
             Kb Kb Kb PCT START/ Kb
TYPE AVAIL USED FREE USED LIMIT RESERVE PRI NAME
dev 1048576 3904 1044672 0% 0 - 1
/dev/vg00/lvol2

this figures are for the working squid2.4. can't give you exakt figures
for 2.5 as we can't use it in production.

this is the general cache info for the working squid2.4.

Connection information for squid:
        Number of clients accessing cache: 0
        Number of HTTP requests received: 127625
        Number of ICP messages received: 0
        Number of ICP messages sent: 0
        Number of queued ICP replies: 0
        Request failure ratio: 0.00%
        HTTP requests per minute: 1174.6
        ICP messages per minute: 0.0
        Select loop called: 896226 times, 7.274 ms avg
Cache information for squid:
        Request Hit Ratios: 5min: 4.8%, 60min: 4.4%
        Byte Hit Ratios: 5min: 8.6%, 60min: 8.9%
        Request Memory Hit Ratios: 5min: 0.6%, 60min: 1.0%
        Request Disk Hit Ratios: 5min: 24.0%, 60min: 22.3%
        Storage Swap size: 471846 KB
        Storage Mem size: 11772 KB
        Mean Object Size: 32.05 KB
        Requests given to unlinkd: 0
Median Service Times (seconds) 5 min 60 min:
        HTTP Requests (All): 0.04519 0.05046
        Cache Misses: 0.04519 0.05331
        Cache Hits: 0.06640 0.07014
        Near Hits: 0.05951 0.07409
        Not-Modified Replies: 0.01309 0.01387
        DNS Lookups: 0.02231 0.03532
        ICP Queries: 0.00000 0.00000
Resource usage for squid:
        UP Time: 6519.125 seconds
        CPU Time: 2708.310 seconds
        CPU Usage: 41.54%
        CPU Usage, 5 minute avg: 21.66%
        CPU Usage, 60 minute avg: 36.44%
        Maximum Resident Size: 23348 KB
        Page faults with physical i/o: 10
Memory usage for squid via mallinfo():
        Total space in arena: 21826 KB
        Ordinary blocks: 17168 KB 7197 blks
        Small blocks: 4008 KB 83200 blks
        Holding blocks: 13 KB 832 blks
        Free Small blocks: 177 KB
        Free Ordinary blocks: 458 KB
        Total in use: 21190 KB 97%
        Total free: 635 KB 3%
Memory accounted for:
        Total accounted: 15106 KB
        memPoolAlloc calls: 16760446
        memPoolFree calls: 16685588
File descriptor usage for squid:
        Maximum number of file descriptors: 1024
        Largest file desc currently in use: 146
        Number of file desc currently in use: 84
        Files queued for open: 0
        Available number of file descriptors: 940
        Reserved number of file descriptors: 100
        Store Disk files open: 0
Internal Data Structures:
         15080 StoreEntries
          1027 StoreEntries with MemObjects
          1024 Hot Object Cache Items
         14720 on-disk objects
Received on Wed May 21 2003 - 09:02:07 MDT