Henrik Nordstrom wrote:

On Wed, 26 Oct 2005, Ken Moberg wrote:

This is a patch to reduce the outage window during a rotate


Outage? Log rotation is supposed to be a fairly quick operation save for the time to write a clean store index (if you Squid is partially swapped out this can take a noticeable amount of time..)

Log rotation is very quick, it all the other stuff that was crammed in there that takes a long time. That's why I wrote a mini rotate that just rotated the basic logs.


+    _db_rotate_log();  /* Rotate cache.log */


If you do this then you need to restart helpers as well... cache.log is conneted to stderr of the helpers.

Do you mean the dns servers or something else?

Ken


Reply via email to