Hi,
I'm currently having a performance issue with Squid 2.6.STABLE18
(running on RHEL4). As I run traffic through the proxy, the memory grows
steadily, and apparently without limit. This increase in memory usage is
coupled with a steadily growing CPU usage, up to a point at which a
single core is saturated (97% usage at ~400MB of RSS). At this point,
the latency of requests increases. When the load is taken off the proxy,
the CPU returns to minimal usage, but the memory usage sticks at the
high water mark.
I should point out that I'm using squid for authentication only (HTTP
digest), not for caching. Consequently, I have maximum_object_size and
maximum_object_size_in_memory both set to 0 in the squid config file. My
understanding is that this should be sufficient to stop squid from
caching.
There's plenty of spare physical RAM on the machine, so it seems
unlikely that it's a memory shortage causing the performance problem. My
interpretation is that something has gotten too large for Squid to
handle but, without object caching, it's not clear to me what that might
be. I would blame the authentication cache, but there's only 2000
different users.
Does anyone have an idea what might be going on, and how to fix it?
Manager info is as follows (taken once the load has been removed).
Connection information for squid:
Number of clients accessing cache: 4
Number of HTTP requests received: 1474327
Number of ICP messages received: 0
Number of ICP messages sent: 0
Number of queued ICP replies: 0
Request failure ratio: 0.00
Average HTTP requests per minute since start: 1864.2
Average ICP messages per minute since start: 0.0
Select loop called: 5783693 times, 8.204 ms avg
Cache information for squid:
Request Hit Ratios: 5min: 0.0%, 60min: 0.0%
Byte Hit Ratios: 5min: -0.0%, 60min: 100.0%
Request Memory Hit Ratios: 5min: 0.0%, 60min: 0.0%
Request Disk Hit Ratios: 5min: 0.0%, 60min: 0.0%
Storage Swap size: 0 KB
Storage Mem size: 104 KB
Mean Object Size: 0.00 KB
Requests given to unlinkd: 0
Median Service Times (seconds) 5 min 60 min:
HTTP Requests (All): 0.00000 0.00000
Cache Misses: 0.00000 0.00000
Cache Hits: 0.00000 0.00000
Near Hits: 0.00000 0.00000
Not-Modified Replies: 0.00000 0.00000
DNS Lookups: 0.00000 0.00000
ICP Queries: 0.00000 0.00000
Resource usage for squid:
UP Time: 47450.907 seconds
CPU Time: 1377.027 seconds
CPU Usage: 2.90%
CPU Usage, 5 minute avg: 0.00%
CPU Usage, 60 minute avg: 0.00%
Process Data Segment Size via sbrk(): 601396 KB
Maximum Resident Size: 0 KB
Page faults with physical i/o: 0
Memory usage for squid via mallinfo():
Total space in arena: 601396 KB
Ordinary blocks: 591340 KB 238750 blks
Small blocks: 0 KB 0 blks
Holding blocks: 255248 KB 5 blks
Free Small blocks: 0 KB
Free Ordinary blocks: 10055 KB
Total in use: 846588 KB 99%
Total free: 10055 KB 1%
Total size: 856644 KB
Memory accounted for:
Total accounted: 98342 KB
memPoolAlloc calls: 275801669
memPoolFree calls: 274326428
File descriptor usage for squid:
Maximum number of file descriptors: 1048576
Largest file desc currently in use: 133
Number of file desc currently in use: 133
Files queued for open: 0
Available number of file descriptors: 1048443
Reserved number of file descriptors: 100
Store Disk files open: 0
IO loop method: epoll
Internal Data Structures:
26 StoreEntries
26 StoreEntries with MemObjects
25 Hot Object Cache Items
0 on-disk objects
squid-
Mark Kent
Principal Software Engineer
Symantec Corporation
www.symantec.com <http://www.symantec.com/>
-----------------------------------------------------
Office: (647) 259-4024
Fax: (647) 259-4018
mark_kent@xxxxxxxxxxxx <mailto:cmark_kent@xxxxxxxxxxxx>
-----------------------------------------------------
______________________________________________________________________
This email has been scanned by the MessageLabs Email Security System.
For more information please visit http://www.messagelabs.com/email
______________________________________________________________________
