[SOLVED] Segmentation fault occurred at

[demades]

Hello,

the following error is repeatly reported in /var/log/messages.
"servidor kernel: grsec: Segmentation fault occurred at xxxxxxxxxxxxxxxx in /opt/zimbra/jdk1.6.0_16/bin/java" where xxxxxxxxxxxxx is a memory position. The parent process sometimes is /opt/zimbra/libexec/zmmtaconfig and sometimes is /opt/zimbra/libexec/zmlogprocess. Anyone knows what is happening?

- zimbra version: 6.0.4 y running on Fedora 11 4G RAM in real hardware.
- i´ve observed the problem the first time two days ago.


Thank you for your help.

[phoenix]

Some idea of operating system, Zimbra version and release would be a good starting point. Please update your with the output of the following command (do not post it in this thread):

zmcontrol -v

How much RAM is on this server and what specification is it? Is it in a VM or real hardware? For how long has this problem been happening?

[shinjuku]

Hello,

I have the same problem on my system.

$ zmcontrol -v
Release 6.0.5_GA_2213.DEBIAN5_64 DEBIAN5_64 FOSS edition.

$ uname -r
2.6.31.5-grsec-xxxx-grs-ipv4-64

The server is real hardware (not VM) with 2GB RAM.

Thank you for your help.

[mperea]

Same problem here:

CentOS 5.4 64 bits with 8 GB RAM.

$ zmcontrol -v
Release 6.0.6_GA_2330.RHEL5_64_20100505212335 CentOS5_64 FOSS edition.

$ uname -r
2.6.32.2-xxxx-grs-ipv4-64

Thx for helping us.

[michelp]

Hello, quite the same problem for me :

Since first fresh install a few days ago, system logs show java/zimbra segmentation fault several (dozens) times a minute in :
/var/log/kern.log
/var/log/syslog

The thing is Zimbra seems to work OK, excepting this.

Anybody an idea to fix it ?

Thanks.

Box is real hardware with Ubuntu-server 8.04 LTS 64 bits, 4GB of RAM :

$ zmcontrol -v
Release 6.0.6_GA_2330.UBUNTU8_64 UBUNTU8_64 FOSS edition.

$ uname -r
2.6.32.2-xxxx-grs-ipv4-64

Typical log sequence in /var/log/kern.log (same in /var/log/syslog ) :

Log from zimbra, looks ok :

May 15 17:40:01 MYSERVER /USR/SBIN/CRON[4824]: (zimbra) CMD (/opt/zimbra/libexec/zmlogprocess > /tmp/logprocess.out 2>&1)
May 15 17:40:01 MYSERVER /USR/SBIN/CRON[4826]: (root) CMD (/usr/local/rtm/bin/rtm 54 > /dev/null 2> /dev/null)
May 15 17:40:01 MYSERVER /USR/SBIN/CRON[4827]: (zimbra) CMD (/opt/zimbra/libexec/zmstatuslog)
May 15 17:40:01 MYSERVER /USR/SBIN/CRON[4825]: (zimbra) CMD (/opt/zimbra/libexec/zmqueuelog)
May 15 17:40:01 MYSERVER /USR/SBIN/CRON[4828]: (zimbra) CMD (/opt/zimbra/libexec/zmcheckduplicatemysqld -e > /dev/null 2>&1)

...and shit happens typically just after it :

May 15 17:40:03 MYSERVER kernel: grsec: Segmentation fault occurred at 0000000000000010 in /opt/zimbra/jdk1.6.0_16/bin/java[java:5312] uid/euid:1000/1000 gid/egid:114/114, parent /opt/zimbra/libexec/zmlogprocess[zmlogprocess:4830] uid/euid:1000/1000 gid/egid:114/114

May 15 17:40:03 MYSERVER kernel: grsec: Segmentation fault occurred at 0000000000000010 in /opt/zimbra/jdk1.6.0_16/bin/java[java:5320] uid/euid:1000/1000 gid/egid:114/114, parent /bin/bash[sh:4987] uid/euid:1000/1000 gid/egid:114/114

Those lines with 'zmlogprocess' = two or more times.

... then several (dozens) lines like this, with 'zmmtaconfig', appearing often each 2 seconds :

May 15 17:40:40 MYSERVER kernel: grsec: Segmentation fault occurred at 0000000000000008 in /opt/zimbra/jdk1.6.0_16/bin/java[java:5963] uid/euid:1000/1000 gid/egid:114/114, parent /opt/zimbra/libexec/zmmtaconfig[zmmtaconfig:13177] uid/euid:1000/1000 gid/egid:114/114

[uxbod]

All the errors posted appear to be a result of using a GRSEC kernel; which I believe is not a stock one delivered with the distro. Akin to RHEL/CENTOS for a successful installation you have to disable SELINUX on those and my assumption is that you would need to do something similar with GRSEC

[mperea]

Changing the kernel to a non GRSEC one did the trick for me.

Thx...