Tek-Tips is the largest IT community on the Internet today!

Members share and learn making Tek-Tips Forums the best source of peer-reviewed technical information on the Internet!

  • Congratulations gkittelson on being selected by the Tek-Tips community for having the most helpful posts in the forums last week. Way to Go!

Where does cldare get local node name?

Status
Not open for further replies.

unixfreak

ISP
Oct 4, 2003
632
GB

As title... Have cloned a cluster and changed ndoenames, network topology etc, but it doesn't even attempt to synchronise/verify just get "cldare: Cannot determine local node name" or something to that effect.

Where does it try to pick this up? I've updated hosts and rhosts files...
 
did you check HACMP's ODM?
I would snapshot cluster configuration and the checked the snapshot (eg. vi) and found if there are any old nodenames or IPs.

Try to correct them, make a copy, remove completly cluster configuration (smitty cm_delete_an_hacmp_cluster_delete.dialog) from the system and then add adjusted snapshot.

once I moved a cluster in that way to different location and it worked fine.
 

Sorry, just realised I'd started this thread, ignore the new one! I've cloned another cluster now and have exactly the same problem! "Removed cluster" before configuring.

I can't actually do a snapshot because it fails running cldare which can't find local node name!

The HACMP ODM is odd, a bunch of .cre files that odmget cna't read. I've grep through them and found nothing interesting...

Hope someone can help
 
Have you looked into /tmp/hacmp.out for more info?

Regards,
Khalid
 

It's not there. Think it only gets generated with the cluster running?

I've looked at other files but nothing interesting

Just been trawling through the ODM but can't really find anything related to a "local node name".


 

Right, in /usr/es/sbin/cluster/utilities the script get_local_nodename is returning nil because the command cllsclstr -N returns nothing.

esrbpltdba:[/]#cllsclstr
ID Name Security Persist

1226764230 esrbplt Standard
esrbpltdba:[/]#cllsclstr -N

esrbpltdba:[/]#

Now the odd thing is that on a working cluster the cllsclstr output is identical (with changed names) so why does it print nothing with the -N flag? Also, HACMPcluster ODM class is correct comparing to a similar cluster... I'm confused.

Anyone know where cllsclstr -N looks?
 

Right, just to not let people hanging...

The cllsclstr -N was inside an if statement and the remaining bit looked for a node name with an IP that matched seomthing setup on an interface with a matching hosts entry. Someone had setup different IP addresses to what I had configured in the HA menus' interface configuration which is why things didn't add up.
 
Status
Not open for further replies.

Part and Inventory Search

Sponsor

Back
Top