Re: [squid-users] storage mem 99% use.

From: Amos Jeffries <squid3_at_treenet.co.nz>
Date: Tue, 09 Jul 2013 21:09:38 +1200

On 9/07/2013 8:17 p.m., Usuario Lista wrote:
> Hello everyone.
> This is the output of my cache settings:
>
>
> ####################################################################
>
> Squid Object Cache: Version 3.3.3
> Start Time: Sun, 16 Jun 2013 08:30:40 GMT
> Current Time: Tue, 09 Jul 2013 05:48:57 GMT
> Connection information for squid:
> Number of clients accessing cache: 10
> Number of HTTP requests received: 3734834
> Number of ICP messages received: 745111
> Number of ICP messages sent: 745514
> Number of queued ICP replies: 0
> Number of HTCP messages received: 0
> Number of HTCP messages sent: 0
> Request failure ratio: 0.00
> Average HTTP requests per minute since start: 113.3
> Average ICP messages per minute since start: 45.2
> Select loop called: 1780628183 times, 1.111 ms avg
> Cache information for squid:
> Hits as % of all requests: 5min: 11.1%, 60min: 11.1%
> Hits as % of bytes sent: 5min: 28.0%, 60min: 29.5%
> Memory hits as % of hit requests: 5min: 38.8%, 60min: 35.8%
> Disk hits as % of hit requests: 5min: 15.0%, 60min: 16.7%
> Storage Swap size: 14155772 KB
> Storage Swap capacity: 90.0% used, 10.0% free
> Storage Mem size: 259448 KB
> Storage Mem capacity: 99.0% used, 1.0% free
> Mean Object Size: 17.49 KB
> Requests given to unlinkd: 614195
> Median Service Times (seconds) 5 min 60 min:
> HTTP Requests (All): 0.00286 0.00286
> Cache Misses: 0.17711 0.15048
> Cache Hits: 0.00000 0.00000
> Near Hits: 0.00598 0.01035
> Not-Modified Replies: 0.00179 0.00000
> DNS Lookups: 0.00000 0.00000
> ICP Queries: 0.00000 0.00000
> Resource usage for squid:
> UP Time: 1977496.987 seconds
> CPU Time: 11355.530 seconds
> CPU Usage: 0.57%
> CPU Usage, 5 minute avg: 1.72%
> CPU Usage, 60 minute avg: 0.84%
> Process Data Segment Size via sbrk(): 687892 KB
> Maximum Resident Size: 2810512 KB
> Page faults with physical i/o: 28916
> Memory usage for squid via mallinfo():
> Total space in arena: 688024 KB
> Ordinary blocks: 664436 KB 3653 blks
> Small blocks: 0 KB 0 blks
> Holding blocks: 37796 KB 11 blks
> Free Small blocks: 0 KB
> Free Ordinary blocks: 23588 KB
> Total in use: 23588 KB 3%
> Total free: 23588 KB 3%
> Total size: 725820 KB
> Memory accounted for:
> Total accounted: 450880 KB 62%
> memPool accounted: 450880 KB 62%
> memPool unaccounted: 274940 KB 38%
> memPoolAlloc calls: 1215726370
> memPoolFree calls: 1228693320
> File descriptor usage for squid:
> Maximum number of file descriptors: 65536
> Largest file desc currently in use: 346
> Number of file desc currently in use: 303
> Files queued for open: 0
> Available number of file descriptors: 65233
> Reserved number of file descriptors: 100
> Store Disk files open: 0
> Internal Data Structures:
> 809371 StoreEntries
> 28483 StoreEntries with MemObjects
> 28481 Hot Object Cache Items
> 809317 on-disk objects
>
> ########################################################################
>
>
> How I can better configure memory usage?
>
> Storage Swap capacity: 90.0% used, 10.0% free
> Storage Mem capacity: 99.0% used, 1.0% free

Just to be clear, there is no problem visible in those numbers. You
allocated Squid some memory for use in a memory cache (cache_mem 256 MB)
Squid is using 99% of it currently. You also allocated some disk for use
in a disk cache (cache_dir 13.5 GB) and Squid is using 90% of that too.

Squid reserves around 10% of the disk space for coping with sudden
floods of new data since disk I/O on delete is slow enough to cause
errors if it has to write to disk before deletions can free up the space.
Memory items are so easily updated or discarded that Squid does not need
to reserve any space as a buffer. Also memory is faster for caching (as
in anything) so Squid prefers to fill it as much as possible with the
most frequently needed objects.

"better" depends on a whole boatload of other factors it is difficult to
guess at. About the only thing you could do better is allocate the
memory cache more memory if the system has plenty to spare or less
memory if it is running out and swapping occasionally.

Amos
Received on Tue Jul 09 2013 - 09:09:56 MDT

This archive was generated by hypermail 2.2.0 : Tue Jul 09 2013 - 12:00:13 MDT