HAMP Problem!

[68ron]

HI

I have problems with changing topology in an active cluster ,actually with synchronizing changes across nodes(EMULATE).

I’ve got the following errors:

In smit screen:

“dsh: 5025-509 db02 rsh had exit code 2
Apr 22 2003 12:16:02Failure occurred while executing cl_emul_cspoc .
Error in executing cl_emul_dare”

and in /tmp/emuhacmp.out

“Error reading status from clsmuxpd
read_cluster_configuration failed
Error in cc_init”

Can somebody help?
Thanks in advanced,

regards

ron

 

[GenFoch]

There is not enough data to understand your problem. How many nodes are in the cluster? How is the cluster set up? cascading or rotating?

From what I have seen, looks like you have a problem with rsh. HA depends on rsh, so if it wont work, HA should fail. Test rsh from each node to the others and make sure its good. This must be done as root, because thats the ID HA uses.

Before you criticize someone, you should walk a mile in their shoes. That way you're a mile away and you have their shoes.

 

[68ron]

HI,

It is a two-node cluster with cascading resource groups(one per node) and now i'm trying to add third node.On one of the active nodes in the cluster I've added a third node to cluster topology, then I've added a adapters for the new node (boot,two standby,service) and also two new serial networks. After the changes made to cluster topology I've try to synchronize cluster topology (EMULATE) but it fails. Cluster veryfication passed without errors and HACMP odm entries are successfully synchronized across nodes but cldare exits with errors mentioned above.

/etc/hosts and /.rhosts files are same on all nodes so i can rlogin in all directions and the connectivity between cluster nodes are established.

Regards,

Ron

 

[DSMARWAY]

hi,

also check the entries in hosts and topology i.e. what entries do you have in your host file for the boot , service and standby i.e. do youhave something like this

serv being the server name

serv_boot
serv_stby
serv_service

or do you have
serv
serv_stby
serv_service


if you type cllsif ( what does it show ) , is your hostname
the same as serv( listed above) because if your hostname is say serv1
and you created your topology boot address as serv1
this normally gets overwritten by the service address so is no longer available , if this is the case change the toplogy entry for boot to serv1_boot , and have an entry in your /etc/hosts file for serv1_boot

Hard to explain , does it make sense ?

 

[68ron]

Hi,

Thanks for advice but I've received the explanation from IBM support which says that this type of error is usually due to filesystem/lv not been seen by all nodes in the cluster and that i should run actual synchronization. I run topology sync without the emulate option and this succeded. It seems like the problem lies in the emulation code , as emulate produces errors , but 'actual' doesnt(

Regards,
Ron