Highlighted

cannot configure vxfen after reboot

Hello,

We move physically a server, and after reboot, we cannot configure vxfen.

# vxfenconfig -c
VXFEN vxfenconfig ERROR V-11-2-1002 Open failed for device: /dev/vxfen with error 2

 

my vxfen.log :

Wed Aug 19 13:17:09 CEST 2015 Invoked vxfen. Starting
Wed Aug 19 13:17:23 CEST 2015 return value from above operation is 1
Wed Aug 19 13:17:23 CEST 2015 output was VXFEN vxfenconfig ERROR V-11-2-1041 Snapshot for this node is different from that of the running cluster.
Log Buffer: 0xffffffffa0c928a0

VXFEN vxfenconfig NOTICE Driver will use customized fencing - mechanism cps
Wed Aug 19 13:17:23 CEST 2015 exiting with 1

 

Engine version 6.0.10.0 

RHEL 6.3

any idea to help me running the vxfen (and the had after ... ) ?

 

7 Replies
Highlighted

Hello, Can you give more info

Hello,

Can you give more info here ... Do you mean you have physically moved one of server in cluster & connected to a different physical location which is connected ?

How many node cluster you have ?

Did you use disk based fencing or cps ?

output of "gabconfig -a " command

cat /etc/vxfenmode    (from all cluster nodes)

cat /etc/vxfendg         (from all cluster nodes)

 

G

Highlighted

Hello, We have physically

Hello,

We have physically moved one of the server of the cluster.

We have 2 nodes.

we configured 2 fencing disk and one cpserver 

From the reboot server :

# gabconfig -a
GAB Port Memberships
===============================================================
Port a gen   a87021 membership 01

From the other (good) server 

t# gabconfig -a
GAB Port Memberships
===============================================================
Port a gen   a87021 membership 01
Port b gen   a87030 membership ;1
Port b gen   a87030    visible 0
Port h gen   a87045 membership ;1
Port h gen   a87045    visible 0

 

 

# cat /etc/vxfenmode (both server are identical)
vxfen_mode=customized
vxfen_mechanism=cps
security=0
fips_mode=0
cps1=[vcs-cptrs.vlan]
port=14250
vxfendg=vxfendg
scsi3_disk_policy=dmp

 

# cat /etc/vxfendg (identical on both server)
vxfendg

 

Thanks for your help.

 

 

 

 

 

Highlighted

Hi, This is how we solved

Hi,

 

This is how we solved the problem.

 

{11} vxfenconfig -c
VXFEN vxfenconfig ERROR V-11-2-1002 Open failed for device: /dev/vxfen with error 2

{12} /etc/init.d/vxfen start
Starting vxfen..
Loaded 2.6.32-358.el6.x86_64 on kernel 2.6.32-358.el6.x86_64
Starting vxfen.. Done
Please see the log file /var/VRTSvcs/log/vxfen/vxfen.log

 

Can you please try this.

Highlighted

Hello Dhawal, we already try

Hello Dhawal,

we already try this :

# /etc/init.d/vxfen stop
VxFEN: Module already unconfigured. Only unloading
# vxfenconfig -c
VXFEN vxfenconfig ERROR V-11-2-1002 Open failed for device: /dev/vxfen with error 2
# /etc/init.d/vxfen start
Starting vxfen..
Loaded 2.6.32-71.el6.x86_64 on kernel 2.6.32-279.el6.x86_64
WARNING:  No modules found for 2.6.32-279.el6.x86_64, using compatible modules for 2.6.32-71.el6.x86_64.
WARNING: All config files need .conf: /etc/modprobe.d/fc-hba.conf.old, it will be ignored in a future release.
Starting vxfen.. Done
Please see the log file /var/VRTSvcs/log/vxfen/vxfen.log

log : 

Thu Aug 20 11:07:12 CEST 2015 executing local_info.sh: end
Thu Aug 20 11:07:12 CEST 2015 calling regular vxfenconfig
Thu Aug 20 11:07:29 CEST 2015 return value from above operation is 1
Thu Aug 20 11:07:29 CEST 2015 output was VXFEN vxfenconfig ERROR V-11-2-1041 Snapshot for this node is different from that of the running cluster.
Log Buffer: 0xffffffffa10158a0

VXFEN vxfenconfig NOTICE Driver will use customized fencing - mechanism cps
Thu Aug 20 11:07:29 CEST 2015 exiting with 1

 

Highlighted

Have you tried unloading the

Have you tried unloading the module from good node & starting it again ... this would need outage to your cluster though ..

Here is what I would suggest ..

On good running node ..

1. Stop had  using hastop.

2. ensure all had / hashadow process are stopped

3. unconfigure fencing.  /etc/init.d/vxfen stop

4. Find fencing module id using lsmod command (if module is still loaded)

5. Unload fencing module (if found loaded in previous step)

6. ensure from lsmod again that fencing module is not loaded

 

On problem node

1. Start fencing first  (/etc/init.d/vxfen start)

On good node

1. Start fencing

 

G

 

Highlighted

The error "Thu Aug 20

The error "Thu Aug 20 11:07:29 CEST 2015 output was VXFEN vxfenconfig ERROR V-11-2-1041 Snapshot for this node is different from that of the running cluster." suggests that on this node where the error occured it is seeing a different disks/cps server than the node where fencing has alread come up.

I would suggest that you check wether the fenbcing disks serial id and the CP server uids match on both the servers.

 

Kindly provide the output of /etc/vxfentab from both the nodes.

 

Regards,

Sudhir

 

Highlighted

Hello, The issue is that

Hello,

The issue is that after reboot of one of them the 2 servers didn't have the same cp server uid's.

the root cause is a change of cp server few weeks ago.

the solution for our case was :

stop the fencing on the server node in error.

stop the cp server.

start the cp server on the old physical one.

check all actifs clusters get connected to.

start the fencing on the node in error

start had 

no error. good.

coffe time Smiley Happy