Re: 1.2 problems on Solaris v2.5 (ss20)

From: Alexander Serkin <als@dont-contact.us>
Date: Thu, 17 Sep 1998 13:15:53 +0400

Alex Rousskov wrote:
>
> What patches did you apply to b24 if any?

none, unfortunately. Do you mean first patch set ?

> For how long did Squid run before stalling?

About two weeks. Until cache was filled out to its limits (~2.7Gb
total).

> Did you --enable-cache-digests?

No

> By "async writes", do you mean --enable-async-io option?
Yes

> Have you tried to shutdown squid with "squid -ks"?
Nope. Just have not read about these command line options :-(

> When killed with "kill -9" would Squid restart immediately or did you have to
> wait for some time?
Have to wait some time... cache.log told that cache_dirs are DIRTY and
made something with them. Then started to read swap logs and claimed
about some disk errors from disk.c. Then stopped answering any queries.
Several times it revived after that and at last the moment comes when i
could not start it at all...

> What was the CPU utilization of Squid after it stalled?
I did not look at it

> Any unusual messages in cache.log?
DIRTY cache dirs and disk.c errors. Something about "assertion".
Unfortunately i lost it after falling back to 1.1.22. logfile_rotate is
set to 1 :-(

> Have you run any recent beta versions before?
No.

--
Alexander
Received on Thu Sep 17 1998 - 01:17:45 MDT

This archive was generated by hypermail pre-2.1.9 : Tue Dec 09 2003 - 16:42:03 MST