[squid-users] Large ACLs and TCP_OUTGOING_ADDRESS

From: Nyamul Hassan <mnhassan_at_usa.net>
Date: Mon, 17 Nov 2008 05:47:57 +0600


I run squid in an ISP scenario. We have got two identically configured
squid caches being load balanced among 4,000 users over a 50 Mbps link. The
system runs quite well, although not without the occassional hiccups. But,
there is a complain from users about not being able to access some websites
because of same external IP. For this, we configured the squid.conf to have
ACLs for different user blocks of /24 and have them mapped through different
external IPs on each of these boxes.

However, not all /24 blocks have the same number of users, and I also have
lots of real IPs still lying unused. I thought about creating different
ACLs for every 5 or 8 users, and then map them to different external IPs.
But, having them distributed in 8 IPs in each group would mean at least 500
separate ACLs and their corresponding TCP_OUTGOING_ADDRESS directives.

My question is, will this affect the performance of squid? Can squid handle

My servers are each running on Core 2 Duo 2.33 GHz, 8 GB of RAM, 5 HDDs
(1x80GB IDE for OS, 4x160GB SATA for cache), total 256GB Cache Store (64GB
on each HDD). One of the server's stats are (taken at a very low user count

Squid Object Cache: Version 2.7.STABLE4
Connection information for squid:
 Number of clients accessing cache: 2281
 Number of HTTP requests received: 46553879
 Number of ICP messages received: 10546598
 Number of ICP messages sent: 10548558
 Number of queued ICP replies: 0
 Request failure ratio: 0.00
 Average HTTP requests per minute since start: 7237.3
 Average ICP messages per minute since start: 3279.5
 Select loop called: 617194936 times, 0.625 ms avg
Cache information for squid:
 Request Hit Ratios: 5min: 30.7%, 60min: 30.2%
 Byte Hit Ratios: 5min: 7.1%, 60min: 5.4%
 Request Memory Hit Ratios: 5min: 20.6%, 60min: 20.5%
 Request Disk Hit Ratios: 5min: 32.3%, 60min: 32.7%
 Storage Swap size: 241785428 KB
 Storage Mem size: 4194120 KB
 Mean Object Size: 35.02 KB
 Requests given to unlinkd: 0
Median Service Times (seconds) 5 min 60 min:
 HTTP Requests (All): 0.72387 0.33943
 Cache Misses: 0.76407 0.55240
 Cache Hits: 0.15048 0.03241
 Near Hits: 0.89858 0.61549
 Not-Modified Replies: 0.04277 0.00286
 DNS Lookups: 0.04433 0.02447
 ICP Queries: 0.03246 0.00037
Resource usage for squid:
 UP Time: 385950.900 seconds
 CPU Time: 51642.987 seconds
 CPU Usage: 13.38%
 CPU Usage, 5 minute avg: 35.09%
 CPU Usage, 60 minute avg: 40.76%
 Process Data Segment Size via sbrk(): 654836 KB
 Maximum Resident Size: 0 KB
 Page faults with physical i/o: 4
Memory usage for squid via mallinfo():
 Total space in arena: -1798084 KB
 Ordinary blocks: 2081758 KB 2399729 blks
 Small blocks: 0 KB 0 blks
 Holding blocks: 35360 KB 8 blks
 Free Small blocks: 0 KB
 Free Ordinary blocks: 314461 KB
 Total in use: -2077186 KB 118%
 Total free: 314461 KB -17%
 Total size: -1762724 KB
Memory accounted for:
 Total accounted: 5839838 KB
 memPoolAlloc calls: 2143236778
 memPoolFree calls: 2113174187
File descriptor usage for squid:
 Maximum number of file descriptors: 65536
 Largest file desc currently in use: 8734
 Number of file desc currently in use: 1413
 Files queued for open: 0
 Available number of file descriptors: 64123
 Reserved number of file descriptors: 100
 Store Disk files open: 62
 IO loop method: epoll
Internal Data Structures:
 6913382 StoreEntries
 324677 StoreEntries with MemObjects
 324228 Hot Object Cache Items
 6905066 on-disk objects

Can anybody advise?

Received on Sun Nov 16 2008 - 23:48:43 MST

This archive was generated by hypermail 2.2.0 : Tue Nov 18 2008 - 12:00:03 MST