I have a web proxy solution (Squid + Dansguardian +75 Ntlm Helpers to
authenticate in the AD) supporting 1500 users.
In the last 3 days the squid started to Die (restart) with the following
errors, about 4 times in the day.
Resume
Warning: all ntlmauthenticator process are busy.
Consider increasing the number of ntlmauthenticator to at least 300
storeDirWriteCleanLogs: Starting ...
Fatal: too many queued ntlmauthenticator requests (376 on 75)
Squid cache : Terminated abnormaly
1 - I was using squid for 5 months and the problem started 3 days ago.
Can a virus or trojan cause this problem?How could I detect it, my squid
has 2 gb logs? Is there any tool or command (grep) that could help me?
- http://www.squid-cache.org/Scripts/
I use (like) calamaris
- MRTG could be used to check traffic going through the SQUID box.
- Check systems logs too , watchout for increase use of system resources
such as swap, for instance due to increased load.
M.
2 - Am I using few authentication helpers for 1500 users? How many
helpers should I use?
3 - Can the size of my logs cause this problem?
Thanks!
