suspected regression in 3.0.28 - applications having problems binding to ports

suspected regression in 3.0.28 - applications having problems binding to ports

[Nikola Ciprich]

[-- Attachment #1: Type: text/plain, Size: 1613 bytes --]

Hi,

I'd like to report following suspicious behaviour of 3.0.27 kernel.

On one of our production boxes, after upgrading from 3.0.26 to 3.0.27,
applications started to behave weird after ~1 day uptime.
Squid started reporting problems binding to port, nagios misbehaved, etc.
I was forced to revert to 3.0.26. We tried to reproduce the problem but were unsuccessful.
After ~1 week without problems with 3.0.26, I tried upgrading this production box to 3.0.27
again to see what happens and the problem appeared again after about one day.

imap started reporting problems:
EMAIL DELIVERY CRITICAL - imap failed: IMAP RECEIVE CRITICAL - Could not connect to A.B.C.D port 993: IO::Socket::INET6 configuration

same for squid:
commBind: Cannot bind socket FD 21 to 0.0.0.0: (98) Address already in use

nagios processes went totally nuts.

There was no other change except for the kernel upgrade.
The box is x86_64 quad CPU 4GB machine, what might be important, we're using
IPv6 as well.
I don't see any suspicious (at least to me) patch between 3.0.26 and 3.0.27.
Since it seems to take quite a long to reproduce, I haven't tried bisect yet, first
wanted to ask - hasn't anyone experienced similar problems?

thanks a lot in advance!

with best regards

nik


-- 
-------------------------------------
Ing. Nikola CIPRICH
LinuxBox.cz, s.r.o.
28. rijna 168, 709 01 Ostrava

tel.:   +420 596 603 142
fax:    +420 596 621 273
mobil:  +420 777 093 799

www.linuxbox.cz

mobil servis: +420 737 238 656
email servis: servis@linuxbox.cz
-------------------------------------

[-- Attachment #2: Type: application/pgp-signature, Size: 198 bytes --]

Re: suspected regression in 3.0.28 - applications having problems binding to ports

[Henrique de Moraes Holschuh]

On Fri, 13 Apr 2012, Nikola Ciprich wrote:
> I'd like to report following suspicious behaviour of 3.0.27 kernel.
> 
> On one of our production boxes, after upgrading from 3.0.26 to 3.0.27,
> applications started to behave weird after ~1 day uptime.
> Squid started reporting problems binding to port, nagios misbehaved, etc.
> I was forced to revert to 3.0.26. We tried to reproduce the problem but were unsuccessful.
> After ~1 week without problems with 3.0.26, I tried upgrading this production box to 3.0.27
> again to see what happens and the problem appeared again after about one day.
> 
> imap started reporting problems:
> EMAIL DELIVERY CRITICAL - imap failed: IMAP RECEIVE CRITICAL - Could not connect to A.B.C.D port 993: IO::Socket::INET6 configuration
> 
> same for squid:
> commBind: Cannot bind socket FD 21 to 0.0.0.0: (98) Address already in use
> 
> nagios processes went totally nuts.
> 
> There was no other change except for the kernel upgrade.
> The box is x86_64 quad CPU 4GB machine, what might be important, we're using
> IPv6 as well.
> I don't see any suspicious (at least to me) patch between 3.0.26 and 3.0.27.
> Since it seems to take quite a long to reproduce, I haven't tried bisect yet, first
> wanted to ask - hasn't anyone experienced similar problems?

Did you test 3.0.28?  Your subject line mentions 3.0.28, which was released
a few hours ago, but the body of the email only mentions a regression in
3.0.27 when compared to 3.0.26...

-- 
  "One disk to rule them all, One disk to find them. One disk to bring
  them all and in the darkness grind them. In the Land of Redmond
  where the shadows lie." -- The Silicon Valley Tarot
  Henrique Holschuh

Re: suspected regression in 3.0.28 - applications having problems binding to ports

[Nikola Ciprich]

[-- Attachment #1: Type: text/plain, Size: 870 bytes --]

> Did you test 3.0.28?  Your subject line mentions 3.0.28, which was released
> a few hours ago, but the body of the email only mentions a regression in
> 3.0.27 when compared to 3.0.26...
ouch, I meant 3.0.27, not 28 :-(
sorry for confusion. I'll be able to test 3.0.28 next wednesday.
BR
nik



> 
> -- 
>   "One disk to rule them all, One disk to find them. One disk to bring
>   them all and in the darkness grind them. In the Land of Redmond
>   where the shadows lie." -- The Silicon Valley Tarot
>   Henrique Holschuh
> 

-- 
-------------------------------------
Ing. Nikola CIPRICH
LinuxBox.cz, s.r.o.
28. rijna 168, 709 01 Ostrava

tel.:   +420 596 603 142
fax:    +420 596 621 273
mobil:  +420 777 093 799

www.linuxbox.cz

mobil servis: +420 737 238 656
email servis: servis@linuxbox.cz
-------------------------------------

[-- Attachment #2: Type: application/pgp-signature, Size: 198 bytes --]

Re: suspected regression in 3.0.28 - applications having problems binding to ports

[Nikola Ciprich]

[-- Attachment #1: Type: text/plain, Size: 1610 bytes --]

Hi,

so I'm sorry for the fuss, the problem actually WAS in nagios, which was
randomly binding to various ports blocking both self and other applications.
Dunno why this always showed up after booting to 3.0.27, but this definitely
is userspace problem.

thanks.

nik


On Fri, Apr 13, 2012 at 09:08:22PM +0200, Nikola Ciprich wrote:
> > Did you test 3.0.28?  Your subject line mentions 3.0.28, which was released
> > a few hours ago, but the body of the email only mentions a regression in
> > 3.0.27 when compared to 3.0.26...
> ouch, I meant 3.0.27, not 28 :-(
> sorry for confusion. I'll be able to test 3.0.28 next wednesday.
> BR
> nik
> 
> 
> 
> > 
> > -- 
> >   "One disk to rule them all, One disk to find them. One disk to bring
> >   them all and in the darkness grind them. In the Land of Redmond
> >   where the shadows lie." -- The Silicon Valley Tarot
> >   Henrique Holschuh
> > 
> 
> -- 
> -------------------------------------
> Ing. Nikola CIPRICH
> LinuxBox.cz, s.r.o.
> 28. rijna 168, 709 01 Ostrava
> 
> tel.:   +420 596 603 142
> fax:    +420 596 621 273
> mobil:  +420 777 093 799
> 
> www.linuxbox.cz
> 
> mobil servis: +420 737 238 656
> email servis: servis@linuxbox.cz
> -------------------------------------



-- 
-------------------------------------
Ing. Nikola CIPRICH
LinuxBox.cz, s.r.o.
28. rijna 168, 709 01 Ostrava

tel.:   +420 596 603 142
fax:    +420 596 621 273
mobil:  +420 777 093 799
www.linuxbox.cz

mobil servis: +420 737 238 656
email servis: servis@linuxbox.cz
-------------------------------------

[-- Attachment #2: Type: application/pgp-signature, Size: 198 bytes --]