We also saw many BAD_MATRIX errors last night.
If the problem was 'wget', shouldn't the snf2check utility detect a corrupt file? Also, we did a manual update yesterday afternoon and there were no 'wget' error messages. The problem got corrected sometime between last night and this morning.
Perhaps though some have had trouble throughout the day.
At the very least the verification on snf2check should be improved to catch this issue. Updating with a bad ruleset creates many problems.
Agreed. I'm looking for some simple ways to do that without changing the rulebase file format. There aren't any simple mechanisms that come to mind. Perhaps there will be no choice but to change the format in order to prevent this possibility.
_M
-----Original Message----- From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] On Behalf Of Pete McNeil Sent: Thursday, March 25, 2004 7:06 PM To: [EMAIL PROTECTED] Subject: Re: [sniffer] Spam storm?
This helps narrow things down. Specifically we know that the rulebase files are not corrupted on the server but during the download. That explains why I haven't been able to recreate a problem in the lab.
I have a suspicion that wget may be failing intermittently. Another customer recently had unexplainable, intermittent issues with wget. They replaced wget with code of their own and have had no further problems.
Can we narrow this down to wget under heavy traffic conditions perhaps?
_M
This E-Mail came from the Message Sniffer mailing list. For information and (un)subscription instructions go to http://www.sortmonster.com/MessageSniffer/Help/Help.html
This E-Mail came from the Message Sniffer mailing list. For information and (un)subscription instructions go to http://www.sortmonster.com/MessageSniffer/Help/Help.html
