Concerned about page fault ratio

From: Steve Snyder <swsnyder@dont-contact.us>
Date: Wed, 15 Mar 2000 15:36:41 -0500 (EST)

I'm concerned about the page fault ratio (pfr) that I'm seeing (Squid
2.2S5, RedHat v6.0 /w kernel 2.2.14).

The full output of "client mgr:info" is below, but here's the short
version:

pfr = Page_faults_with_physical_i/o / Number_of_HTTP_requests_received

0.47 = 9112 / 19368

(The fault and requests numbers above are low due to my having
rebooted recently, but the ratio calculated is representative of what
I've been seeing.)

This ratio is much higher than the examples I see in the Squid FAQ. I
don't any reason for Squid to have to be swapped out. I've got plenty
of RAM for the programs running on this box and the swap partition is
barely used.

Should I be concerned about the page fault ratio or do these numbers
seem OK?

Thank you.

------------------------------------------

HTTP/1.0 200 OK
Server: Squid/2.2.STABLE5
Mime-Version: 1.0
Date: Wed, 15 Mar 2000 20:21:08 GMT
Content-Type: text/plain
Expires: Wed, 15 Mar 2000 20:21:08 GMT
Last-Modified: Wed, 15 Mar 2000 20:21:08 GMT
X-Cache: MISS from proxy.snydernet.lan
Proxy-Connection: close

Squid Object Cache: Version 2.2.STABLE5
Start Time: Thu, 09 Mar 2000 17:45:25 GMT
Current Time: Wed, 15 Mar 2000 20:21:08 GMT
Connection information for squid:
        Number of clients accessing cache: 5
        Number of HTTP requests received: 19368
        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: 2.2
        ICP messages per minute: 0.0
        Select loop called: 1478397 times, 356.969 ms avg
Cache information for squid:
        Request Hit Ratios: 5min: 0.0%, 60min: 61.0%
        Byte Hit Ratios: 5min: -0.0%, 60min: 2.0%
        Storage Swap size: 237198 KB
        Storage Mem size: 9928 KB
        Storage LRU Expiration Age: 102.02 days
        Mean Object Size: 8.17 KB
        Requests given to unlinkd: 2951
Median Service Times (seconds) 5 min 60 min:
        HTTP Requests (All): 0.17711 0.05046
        Cache Misses: 0.17711 0.37825
        Cache Hits: 0.00000 0.00405
        Near Hits: 0.00000 0.27332
        Not-Modified Replies: 0.00000 0.00721
        DNS Lookups: 0.00000 0.00094
        ICP Queries: 0.00000 0.00000
Resource usage for squid:
        UP Time: 527742.522 seconds
        CPU Time: 71.230 seconds
        CPU Usage: 0.01%
        CPU Usage, 5 minute avg: 0.00%
        CPU Usage, 60 minute avg: 0.05%
        Maximum Resident Size: 0 KB
        Page faults with physical i/o: 9112
Memory usage for squid via mallinfo():
        Total space in arena: 17246 KB
        Ordinary blocks: 17222 KB 28 blks
        Small blocks: 0 KB 0 blks
        Holding blocks: 176 KB 1 blks
        Free Small blocks: 0 KB
        Free Ordinary blocks: 23 KB
        Total in use: 17398 KB 101%
        Total free: 23 KB 0%
Memory accounted for:
        Total accounted: 14025 KB
File descriptor usage for squid:
        Maximum number of file descriptors: 1024
        Largest file desc currently in use: 15
        Number of file desc currently in use: 16
        Files queued for open: 0
        Available number of file descriptors: 1008
        Reserved number of file descriptors: 100
        Store Disk files open: 0
Internal Data Structures:
         29055 StoreEntries
          2150 StoreEntries with MemObjects
          2149 Hot Object Cache Items
         29021 Filemap bits set
         29021 on-disk objects

*** Steve Snyder ***
Received on Wed Mar 15 2000 - 13:40:11 MST

This archive was generated by hypermail pre-2.1.9 : Tue Dec 09 2003 - 16:52:10 MST