Messaging Gateway

ALLCON,

For the past week now there has been some odd behavior from out BM Server, it keeps shutting down and coming back on-line.

These are the messages we have received so far:

Fri 1/28/2011 3:20 AM

bmserver crashed on signal 6 on bmgate.domain.com exit code: 0x0086

Program output:

*** glibc detected *** corrupted double-linked list: 0x9248c620 ***

Storing data in /data/scanner/jobs/bmserver/2011.01.28-03.18.16

Sun 1/30/2011 2:19 AM

conduit crashed on signal 6 on bmgate.domain.com exit code: 0x0086

Program output:

*** glibc detected *** double free or corruption (!prev): 0x0929f668 ***

Storing data in /data/scanner/jobs/conduit/2011.01.30-02.19.08

Tue 2/1/2011 6:22 PM

bmserver crashed on signal 11 on bmgate.domain.com exit code: 0x008B

Storing data in /data/scanner/jobs/bmserver/2011.02.01-18.20.38

Tue 2/1/2011 8:22 PM

bmserver crashed on signal 6 on bmgate.domain.com exit code: 0x0086

Program output:

*** glibc detected *** free(): invalid next size (normal): 0x0853c4e8 ***

Storing data in /data/scanner/jobs/bmserver/2011.02.01-20.21.20

There has not been a crash since last Tuesday, however this does not mean that it will not happen again.

Any insight on this would be most helpful.

Thank you for your time in advance.

Can you tell us a bit more about the environment? What version of SBG? Are you checking SPF records? if so, for all domains or just specfic?

We are running SBG 9.0.2. It's a small setup, 50 - 100 users. I have not had the chance to look at the SPF records as I've been deploying new machines and fixing issues. We just had a massive crash today of the brightmail server. We just had 9 mta crashed on signal 11 errors, 1 conduit crashed on signal 11 errors, and 2 improper shutdown/startup errors all within a 10 minute span. The issues stopped when I posted this last Tuesday and now we are getting massive issues today.

Have you opened up a support case on this issue?  We'd love to get more details on the crashes, but the best process is to start with a support case.  I'll track the case if you send me the case number.

Angelos Kottas

Brightmail Gateway Product Manager

Server installed today and every 10 minutes we receive same error:

Subject:   Cron <root@LVRELP01> /usr/lib/sa/sa1 1 1

Message: *** glibc detected *** double free or corruption (!prev): 0x088d59f8 ***

Looks like this error is generated when Live Update starts (default 10 minutes)

I can send diagnostics files if you want them.

No more errors here :-)

Here it was a NTP problem. The server could not connect to a private NTP server. After changing the settings to a internet NTP pool the problem was solved.