OpenVZ Forum


Home » General » Support » "oops in connect_init" in kernel ring buffer
"oops in connect_init" in kernel ring buffer [message #8054] Sat, 04 November 2006 07:36 Go to next message
rickb is currently offline  rickb
Messages: 368
Registered: October 2006
Senior Member
Hi, I booted into 2.6.18 with patch patch-ovz028test002-combined and config kernel-2.6.18-028test002-i686-enterprise.config.ovz on a dual core dual cpu xeon server. After 26 minutes of uptime see 74 of these entries in dmesg:

Oops in connect_init! tp->advmss=0, dst->advmss=16396
dst: pmtu=16436, advmss=16396
sk->state=2, tp: ack.rcv_mss=0, mss_cache=524, advmss=0, user_mss=0


and a 48 of these:

Oops in connect_init! tp->advmss=0, dst->advmss=1460
dst: pmtu=1500, advmss=1460
sk->state=2, tp: ack.rcv_mss=0, mss_cache=524, advmss=0, user_mss=0


The server seems to be running fine, what does it mean and is it anything to worry about?

Thanks!
Rick


-------------
Common Terms I post with: http://wiki.openvz.org/Category:Definitions

UBC. Learn it, love it, live it: http://wiki.openvz.org/Proc/user_beancounters

[Updated on: Sat, 04 November 2006 07:37]

Report message to a moderator

Another oddity with kernel-smp-2.6.18-ovz028test002.1 [message #8070 is a reply to message #8054] Mon, 06 November 2006 20:15 Go to previous messageGo to next message
dowdle is currently offline  dowdle
Messages: 261
Registered: December 2005
Location: Bozeman, Montana
Senior Member
Greetings,

I also see those opps message in my /var/log/messages... but I'll just pile on here... since this topic is about the new 2.6.18 kernel. Perhaps I'll end up filing a bugzilla report.

On the physical host, it shows an odd process:

[root@new ~]# w
 15:08:16 up 1 day, 21:20,  2 users,  load average: 0.08, 0.25, 0.25
USER     TTY      FROM              LOGIN@   IDLE   JCPU   PCPU WHAT
root     pts/0    65-100-77-107.bl Sat20   26:13m  2.09s  2.03s tail -f /home/cvs/server-root/mint/CVSROOT/history
root     pts/1    scott.coe.montan 15:00    0.00s  0.04s  0.01s w


Further investigation shows:

[root@new ~]# ps auxwww | grep cvs
501      18877  0.0  0.0  1848  560 /vz/root/104/dev/pts/0 S+ Nov05   0:02 tail -f /home/cvs/server-root/mint/CVSROOT/history


Noticed that w shows the process being run on the physical host but ps actually shows it is being run on vps 104.

More info:

[root@new ~]# vzctl enter 104
entered into VE 104
-bash-3.00# w
 15:10:15 up 1 day, 21:18,  1 user,  load average: 0.00, 0.05, 0.04
USER     TTY      FROM              LOGIN@   IDLE   JCPU   PCPU WHAT
frank    pts/0    wh58-508.st.uni- Sat18   36:30m  2.03s  2.03s tail -f /home/cvs/server-root/mint/CVSROOT/history


So, it appears that w on the physical host is confused and thinks root from host X is running the tail... when in reality, it is a completely different user, sshed in from a completely different host on a VPS rather than the physical host.

Weird, huh?


--
TYL, Scott Dowdle
Belgrade, Montana, USA
Re: Another oddity with kernel-smp-2.6.18-ovz028test002.1 [message #8074 is a reply to message #8070] Tue, 07 November 2006 02:19 Go to previous messageGo to next message
rickb is currently offline  rickb
Messages: 368
Registered: October 2006
Senior Member
Showing vps user activity in the HN (under the output of w) has been in vz and openvz for as long as I can remember. It only happens when you enter and exit a vps with "vzctl enter" and then run w on the HN. I wouldn't say it happens 100% of the time though.

Rick


-------------
Common Terms I post with: http://wiki.openvz.org/Category:Definitions

UBC. Learn it, love it, live it: http://wiki.openvz.org/Proc/user_beancounters
Re: Another oddity with kernel-smp-2.6.18-ovz028test002.1 [message #8078 is a reply to message #8070] Tue, 07 November 2006 08:08 Go to previous message
Vasily Tarasov is currently offline  Vasily Tarasov
Messages: 1345
Registered: January 2006
Senior Member
There're problems with w. You can search in forum about it. Generally it happens due to two reasons:
1) algorithm, that w uses to get information (it uses /proc/<pid>).
2) in VE0 you can see all processes -> all /proc/<pid> entries.

As concerns
Quote:

Oops in connect_init! tp->advmss=0, dst->advmss=16396
dst: pmtu=16436, advmss=16396
sk->state=2, tp: ack.rcv_mss=0, mss_cache=524, advmss=0, user_mss=0

Please, fille the bug.

Thanks,
Vasily.
Previous Topic: Kernel config for 2.6.18
Next Topic: *SOLVED* Kernel Panic!?
Goto Forum:
  


Current Time: Fri Nov 15 07:05:04 GMT 2024

Total time taken to generate the page: 0.03159 seconds