[squid-users] Fine tuning of Squid and OS configuration for handling more loads

[Alex Rousskov]

On 4/8/22 05:56, pvs wrote:

 > sudo /usr/squid/5.2/sbin/squid -f /etc/squid/squid.conf

Next time, try starting Squid from the console/terminal using the 
following command:

     sudo /usr/squid/5.2/sbin/squid -Nd1 -f /etc/squid/squid.conf

This is _not_ the right way to start Squid in production, but it is a 
good way to test whether you can get some error messages on the console 
after Squid dies during this _test_ run.


 > I wanted to share cache.log file, but the size is too big.

You can always share a link to a (compressed) log file or at least 
copy-paste the last ~100 lines written just _before Squid exited_. The 
exiting lines, if any, would be printed above the Squid startup lines 
similar to these:

 > 2022/04/07 18:43:38.457| Created PID file (.../squid.pid)
 > 2022/04/07 18:43:38.459| Set Current Directory to /usr/local/squid/var
 > 2022/04/07 18:43:38.459| Starting Squid Cache version ...

The other information you have provided was useful but not sufficient.


Please keep this conversation on the squid-users list instead of 
responding to me personally.


Thank you,

Alex.


 > On 07-04-2022 19:09, Alex Rousskov wrote:
 >> On 4/7/22 00:15, pvs wrote:
 >>
 >>> On the 3rd day morning all of a sudden squid process stopped, after
 >>> restarting it worked 10 minutes and again stopped, this process
 >>> repeated and finally we removed this squid from production network.
 >>
 >>> We tried to analyze the syslog file, squid access/cache log files,
 >>> not got any concrete clues.
 >>>
 >>> Any clues/ideas where i can look for additional info for the squid
 >>> process all of a sudden dying.
 >>
 >> How do you start Squid?
 >>
 >> How many "sbin/squid" or equivalent processes are running when
 >> everything is working OK?
 >>
 >> You may want to share "ps aux | fgrep squid" or similar output after
 >> starting Squid the usual way.
 >>
 >>
 >> If Squid knows that it is dying, then cache.log usually has a record
 >> of the cause of death. Squid will often known that it is dying, but
 >> there are exceptions. Squid will know less if you start it with a -C
 >> command line option.
 >>
 >> If Squid does not know that it is dying, then whatever scripts/system
 >> you used to start Squid may have a record of it. Usually, those
 >> messages go into syslog.
 >>
 >> I know you said you looked in cache.log and syslog, but perhaps there
 >> are some clues there that you consider unimportant. If anything is
 >> logged at the time of death, consider sharing those records.
 >>
 >>
 >> If there are really no traces anywhere, then you are doing something
 >> very unusual, but you should still be able to triage the issue if you
 >> start Squid by hand from the terminal (add "-d1" if you are not using
 >> that option already) and watch the console output at the time of
 >> death. You will also be able to collect the exit code of the top-level
 >> squid process this way -- another potential clue.
 >>
 >>
 >> HTH,
 >>
 >> Alex.
 >
 > --
 > Regards,
 >
 > पं. विष्णु शंकर		P. Vishnu Sankar
 > टीम लीडर                Team Leader-Network Operations
 > सी-डॉट                  C-DOT
 > इलैक्ट्रॉनिक्स सिटी फेज़ I    Electronics City Phase I
 > होसूर रोड बेंगलूरु          Hosur Road Bengaluru – 560100
 > फोन  Ph                91 80 25119466
 > ------------------------------------------------------