DIA3202C The TCP/IP call "setsockopt(TCP_KEEPCNT)" returned an errno="22".

Discussion:

(too old to reply)

b***@gmail.com

2017-12-21 13:13:36 UTC

Hi All -

Running DB2 10.5 FP 8 on AIX 7.1.

I'm getting this error every few minutes

2017-12-21-05.59.50.374425-360 E154360A429 LEVEL: Error
PID : 51839300 TID : 2314 PROC : db2sysc 0
INSTANCE: db2inst1 NODE : 000
HOSTNAME: p1proddb.vanhoof.com
EDUID : 2314 EDUNAME: db2tcpcm 0
FUNCTION: DB2 UDB, common communication, sqlcctcpconnmgr_child, probe:2819
MESSAGE : DIA3202C The TCP/IP call "setsockopt(TCP_KEEPCNT)" returned an
errno="22".

Nothing has changed on the DB-side but I am told by one of the network guys that a windows NFS server 'moved' to a new location or was replaced by another server and then this started happening.

The IBM documentation sez it isn't a DB2 issue per-se but the error msgs are becoming tiresome and annoying.

Any ideas here? Anybody see this before?

-Bruce

Jerry Stuckle

2017-12-21 14:40:00 UTC

Permalink

Post by b***@gmail.com
Hi All -
Running DB2 10.5 FP 8 on AIX 7.1.
I'm getting this error every few minutes
2017-12-21-05.59.50.374425-360 E154360A429 LEVEL: Error
PID : 51839300 TID : 2314 PROC : db2sysc 0
INSTANCE: db2inst1 NODE : 000
HOSTNAME: p1proddb.vanhoof.com
EDUID : 2314 EDUNAME: db2tcpcm 0
FUNCTION: DB2 UDB, common communication, sqlcctcpconnmgr_child, probe:2819
MESSAGE : DIA3202C The TCP/IP call "setsockopt(TCP_KEEPCNT)" returned an
errno="22".
Nothing has changed on the DB-side but I am told by one of the network guys that a windows NFS server 'moved' to a new location or was replaced by another server and then this started happening.
The IBM documentation sez it isn't a DB2 issue per-se but the error msgs are becoming tiresome and annoying.
Any ideas here? Anybody see this before?
-Bruce

As the IBM documentation states:

"Network and System Admins should investigate their networks and systems
to determine underlying root cause for the TCP troubles."

This is not a DB2 problem.

--
==================
Remove the "x" from my email address
Jerry Stuckle
***@attglobal.net
==================

b***@gmail.com

2017-12-22 13:08:18 UTC

Permalink

Post by Jerry Stuckle

"Network and System Admins should investigate their networks and systems
to determine underlying root cause for the TCP troubles."
This is not a DB2 problem.
--
==================
Remove the "x" from my email address
Jerry Stuckle
=============

Yes I understand that this isn't necessarily a DB2 Issue but the fact remains we're getting 100s of these a day and the volume of these is such its potentially masking out other issues.

So, it may not be DB2 caused but it affects DB2 and there doesn't seem to be a good reason or fix.

Jerry Stuckle

2017-12-22 14:32:13 UTC

Permalink

Post by b***@gmail.com

Post by Jerry Stuckle

Yes I understand that this isn't necessarily a DB2 Issue but the fact remains we're getting 100s of these a day and the volume of these is such its potentially masking out other issues.
So, it may not be DB2 caused but it affects DB2 and there doesn't seem to be a good reason or fix.

This is NOT a DB2 issue and the fact remains if you want to fix it you
need to proceed in the proper way. And IBM told you how to do it.

There is no way to troubleshoot this from DB2 - as IBM explained (and
explained why).

--
==================
Remove the "x" from my email address
Jerry Stuckle
***@attglobal.net
==================

Peter H. Coffin

2017-12-29 20:43:56 UTC

Permalink

Post by b***@gmail.com

^^^^^^^^

Post by b***@gmail.com

Post by b***@gmail.com
FUNCTION: DB2 UDB, common communication, sqlcctcpconnmgr_child, probe:2819
MESSAGE : DIA3202C The TCP/IP call "setsockopt(TCP_KEEPCNT)" returned an
errno="22".

[..]

Post by b***@gmail.com
Yes I understand that this isn't necessarily a DB2 Issue but the fact
remains we're getting 100s of these a day and the volume of these is
such its potentially masking out other issues.
So, it may not be DB2 caused but it affects DB2 and there doesn't seem
to be a good reason or fix.

db2tcpcm is a network listening process. Something is connecting to
p1proddb.vanhoof.com then the connection is dropping. Find nearby log
messages to determine what is connecting. Find owner of offending
machine. Proceed with beatings until owner of machine stops the
connection attempts. Or block the other machine from connecting to the
DB host at all. It's not a DB2 issue.

--
Compared to system administration, being cursed forever is a step up.
-- Paul Tomko