[squid-users] PLEASE HELP: Large cache deployment and squid

PLEASE HELP

I have been trying to post a questoin about very large cache to the list but
have not succeeded.

Anyways, I have the following situation, and don't know if anybody has faced
it
as well.

2 * 1GHz PIII
Intel Server Motherboard (ServerWorks PCI compiled in the kernel.)
1GB RAM
cache_mem = 2MB
pagecache = 2 15 75
SCSI HDD
cache_dir = 3 * 13GB on different disks
linux kernel = 2.4.2-2 smp (recompiled to remove all unnecessary stuff and )
squid version = 2.4STABLE1 (from redhat rawhide) (had to install openssl as
well)

After 12 hours of use even 150+ request / minute (yes minutes) squid takes
up 70%+ CPU.
It has also happened to me with 2.3STABLE4 (from default rh7.1 linux cd)
kreiserfsd taking up 10% cpu at times

Let me know if you have faced this situation as well.

memory cache goes to 555MB+ that's what vmstat shows. squid takes up ~20MB
as shown by top

Strace showed me the following after the 12 hour check (poll %time is
surprising)

% time seconds usecs/call calls errors syscall
------ ----------- ----------- --------- --------- ----------------
-0.05 1052.535622 1587 663017 106 write
-0.03 699.529842 1151 607949 121402 read
-0.01 301.1657568 10431 29014 29014 connect
-0.01 291.819954 1990 146680 close
-0.00 12.1088647 4 3583041 gettimeofday
-0.00 1.459970 15 97439 45321 recvfrom
-0.00 1.406372 48 29039 socket
-0.00 0.1175930 23 52143 sendto
-0.00 0.893687 403 2215 dup2
-0.00 0.767492 26 29039 setsockopt
-0.00 0.701222 2 440545 fcntl64
-0.00 0.608259 18432 33 stat64
-0.00 0.588163 7 87557 41100 accept
-0.00 0.373924 407 918 getrusage
-0.00 0.328045 109348 3 statfs
-0.00 0.328029 109343 3 mmap2
-0.00 0.328027 109342 3 munmap
-0.00 0.291677 863 338 brk
-0.00 -1.1256203 6 46457 getsockname
0.00 -1.906077 -2 29383 705 getsockopt
0.00 -1.762877 -11 19811 lseek
0.00 -2.702382 -23 54514 fstat64
0.00 -6.888973 -715 7138 getdents64
0.00 -24.561665 -329 71049 2 open
100.11 -2088636.745941 -440 2907882 poll
------ ----------- ----------- --------- --------- ----------------
100.00 -2086297.706548 8905210 237650 total

I have also realised that the pagecache has eaten up 575MB RAM and that's
the only difference after rebooting the server.

After a reboot, the response times are excellent, but in few hours when the
pagecache builds up i.e. shows up in vmstat as "memory cache", the
performance goes down.

btw reiserfs is 3.6.25 standard that comes with redhat. I have not installed
anything other kernel patches.

Does anybody have similar problems? I have been struggling for a week and no
signs of better response.

Is there anything that can let me know what is the problem on the server?
Squid takes up 100% if the server is reboot.

I even upgraded Squid to 2.4S1 and linux 2.4.6. Anyways I can find out what
squid is doing when it is taking up all this power.

All documents I read says Squid does not require more processing power, yet
I gave it 2 x 1GHz. Is that a problem?

I hope this helps you find what is the problem.

Regards,

Mitesh Choksi
Received on Wed Jul 25 2001 - 00:10:12 MDT