MC/SG problem.

[meng_1]

Hi all,
My eniv is. N4000 HP-UX 11.0 OS with MC/SG(11.07) the database is ORACLE 8.0.5
my problem is the that the application automatic swith form host1 to host2. I found some erroe is syslog is that.

tjbs2 tftpd[1375]: Timeout (no requests in 10 minutes)

I have attached the syslog, pls help me check.
I have installed the PHSS_27158 for MC/SG 11.07.

thanks in advance

[Jeff Schussele]

Hi Meng,

Unless you have some sort of tftp process tied into the pkg that's switching. I doubt that's the cause.

You need to look in the pkg log itself for clues. It's usually in /etc/cmcluster/pkg_name/pk_name.log

I suspect you've probably got a heartbeat timing out for whatever reason & that's causing it. But it could be the pkg itself or any number of things.

Good Hunting,
Jeff

[meng_1]

Hi Jeff,

I thank you so much for your respond,
you r correct, that have some problem on the heartbeat line, I have chage the heartbeat cable the problem is still appear.
I will attach the pkg log later.

U know the heartbeat lan port is on the core/IO card. Do U think that card have problem? this is N4000 server.

talk to U later.

[Jeff Schussele · Accepted solution]

Hi,

Yes that's entirely possible. So is the possibility that it's the cable or hub/switch port.

But important questions are also:
1) Is the hearbeat on a public network?
2) What's the timeout value defined for the heartbeats?
3) If it is on a public network & you're using the default value of 2 seconds, you either need to move to a private heartbeat-only network or at least increase that timeout to 8 - 10 seconds.
You may be timing out due to network congestion.

Rgds,
Jeff

[meng_1]

Hi Jeff,

Now I attached the pkg log as ZIP file.
I think there maybe have no useful information in the log file.

and thanks a again.

mxz