topic Re: heartbeat status reporting down? in Operating System - Linux

heartbeat status reporting down?

C Lamb — Tue, 05 May 2009 15:31:30 GMT

I havent seen this in the Forum so I thought I'd ask. We have built several 2 node clusters and are seeing something strange. on several of our fully functional clusters the cmviewcl -v is reporting the status of both heartbeat interfaces as down. The clusters are fine. why is this? shouldn't they report as up? they do on some of the other clusters.

Re: heartbeat status reporting down?

Steven E. Protter — Tue, 05 May 2009 19:03:39 GMT

Shalom,

Network collision/congestion can cause this condition.

Can I actually see the cmviewcl -v output.

SEP

Re: heartbeat status reporting down?

C Lamb — Tue, 05 May 2009 19:05:50 GMT

Here's the output,
[root@maildb01a maildb01]# cmviewcl -v

CLUSTER STATUS
maildb01 up

NODE STATUS STATE
maildb01a up running

Cluster_Lock_LUN:
DEVICE STATUS
/dev/sdj1 up

Network_Parameters:
INTERFACE STATUS NAME
PRIMARY up eth0
PRIMARY down eth4
PRIMARY down eth2

PACKAGE STATUS STATE AUTO_RUN NODE
maildb01 up running enabled maildb01a

Policy_Parameters:
POLICY_NAME CONFIGURED_VALUE
Failover configured_node
Failback manual

Script_Parameters:
ITEM STATUS MAX_RESTARTS RESTARTS NAME
Subnet up 10.10.120.0

Node_Switching_Parameters:
NODE_TYPE STATUS SWITCHING NAME
Primary up enabled maildb01a (current)
Alternate up enabled maildb01b

NODE STATUS STATE
maildb01b up running

Cluster_Lock_LUN:
DEVICE STATUS
/dev/sdj1 up

Network_Parameters:
INTERFACE STATUS NAME
PRIMARY up eth0
PRIMARY down eth4
PRIMARY down eth2
[root@maildb01a maildb01]#

Re: heartbeat status reporting down?

C Lamb — Tue, 05 May 2009 19:08:07 GMT

and here is the other cluster, and I know the package is down

cmviewcl -v

CLUSTER STATUS
cache_cluster up

NODE STATUS STATE
lnx-oradb01 up running

Cluster_Lock_LUN:
DEVICE STATUS
/dev/sdb1 up

Network_Parameters:
INTERFACE STATUS NAME
PRIMARY up eth0
PRIMARY up eth2
PRIMARY up eth3

NODE STATUS STATE
lnx-oradb02 up running

Cluster_Lock_LUN:
DEVICE STATUS
/dev/sdq1 up

Network_Parameters:
INTERFACE STATUS NAME
PRIMARY up eth0
PRIMARY up eth2
PRIMARY up eth3

UNOWNED_PACKAGES

PACKAGE STATUS STATE AUTO_RUN NODE
cache1p2p down halted disabled unowned

Policy_Parameters:
POLICY_NAME CONFIGURED_VALUE
Failover configured_node
Failback manual

Script_Parameters:
ITEM STATUS NODE_NAME NAME
Subnet up lnx-oradb01 10.10.132.0
Subnet up lnx-oradb02 10.10.132.0

Node_Switching_Parameters:
NODE_TYPE STATUS SWITCHING NAME
Primary up enabled lnx-oradb01
Alternate up enabled lnx-oradb02

Re: heartbeat status reporting down?

John Bigg — Wed, 06 May 2009 05:55:28 GMT

If you are running RH 5.2 with lans which use the e1000e driver then this will be the cause of the problem. There is a bug in the e1000e driver which causes it to report an incorrect status. The solution here would be to upgrade to RH 5.3 which has the driver bug fixed.

Re: heartbeat status reporting down?

C Lamb — Wed, 06 May 2009 13:59:59 GMT

we are on RH 5.3 and both are using Intel Pro/1000. I do think your right though, its just a reporting issue, the cluster performs properly. do you happen to know the exact drivers involved?