inetd going down

[V.Tamilvanan]

Hi all,
The inetd daemon on rp7410 HP-UX 11i (new) servers are dying automatically with the syslog error "inetd[603]: Going down on signal 11". It also creates a core file when it dies.
Can someone let me know whether any patch needs to be installed.

Thanks & regards,
V.Tamilvanan

[avsrini]

Hi Tamil,
Check the following.

PROBLEM
The following version of inetd(1M)

version PHNE_21835

core dumps after receiving a signal 11 SIGSEV - segmentation violation.

CONFIGURATION
Operating System - HP-UX
Version - 11.0
Subsystem - inetd
RESOLUTION
There are currently two causes identified for this problem:

1. a code defect within inetd

The core may occur if the rpc.ttdbserver service is enabled
in /etc/inetd.conf.

This problem is currently fixed in the following inetd
Site Specific patches:

PHNE_26365
(for HP-UX 11.0)

PHNE_26366
(for HP-UX 11.11)

Note: To determine whether or not either of these Site Specific
patches may be appropriate for your environment, please
contact your local HP Support Center.

(HP Engineers who encounter this problem before a GR
is available are asked to open a new case with the
HPS-SE [WTEC] EC NET UNIX IS team.)

The following General Patch is planned for release in
November 2002:

PHNE_23708
( HP-UX 11.0 inetd GR patch)


2. a code defect in /lib/pa20_64/dld.sl (dynamic loader)

This problem was originally fixed in the following General
Release patches:

PHSS_23440
11.00 ld(1) and linker tools cumulative patch

(replaced by PHSS_26559)

PHSS_23441
11.11 ld(1) and linker tools cumulative patch

(replaced by PHSS_26560)


Note: These patches, as with any patches, may be superseded.
Please check for the latest patches at the IT Resource
Center (ITRC) at the following web site:

http://www.itrc.hp.com

Srini.

[Michael Tully]

If this is a new server, which you have indicated, it could be that you haven't increased any kernel parameters since your build. A signal 11 does not necessarily mean that there is something wrong with the 'inetd' binary or other code.

# grep 11 /usr/include/sys/errno.h
* @(#)B.11.11_LR common/sys/errno.h $Revision: $
#define EAGAIN 11 /* No more processes */

Have a look at 'nproc' any it's dependencies.
http://www.docs.hp.com/hpux/onlinedocs/939/KCParms/KCparams.OverviewAll.html

[U.SivaKumar_2]

Hi,

signal 11 represents the segmentation fault. when the executable tries to access memory locations that haven't been allocated for the it's use.

Things to check .

1) patches for bugs in inetd
2) your free memory
3)Permission of /usr/sbin/inetd should be
-r-xr--r-- 1 bin bin

regards,
U.SivaKumar

[Ravi_8]

Hi,

The reason that inetd is dying is becuase it is accessing an invalid portion of
memory. Signal 11 is a memory segment violation. A debug version of inetd has
been created to further investigate the cause of this issue and is available by
opening a call with the response center.

I have installed these patches ( HP-UX 11.0)
PHSS_25787
PHSS_25199
PHCO_24838

According to HP support that should help.
Pls find out equivalent for 11i.
Hope it helps.