Unsolved
This post is more than 5 years old
7 Posts
0
2702
June 10th, 2015 19:00
Failover MDM
Hi,
Any idea how to enable the failover MDM ? I saw the mdm_failover binary on MDM server, managed to bring it up but when I did a simulation with secondary MDM goes down, the failover didn't do anything. Can't find the manual either for mdm_failover.
Regards,
No Events found!
dondy1
7 Posts
0
June 11th, 2015 05:00
Hi,
Thanks for the reply.
Yes, I meant the mdm_failover explicitly. However, I can't access link that you've given me
Not authorized
Regards,
echolaughmk
2 Intern
•
522 Posts
0
June 11th, 2015 05:00
Hi,
If the goal is to switch the MDM to the Secondary MDM, you can do this through the CLI:
If you are talking about MDM_failover explicitly, then that is a little more involved given additional standby or tie-breaker MDMs that can be configured and this is a good manual that discusses that scenario:
https://support.emc.com/docu56714_Automatic-MDM-Cluster-Node-Replacement-Technical-Note.pdf?language=en_US
HTH
dondy1
7 Posts
0
June 11th, 2015 06:00
Hi,
Already follow you
need to be followed back so I can send you private message
echolaughmk
2 Intern
•
522 Posts
0
June 11th, 2015 06:00
That might be due to the credentials associated with your account. Send me a direct message offline and I can send it to you.
chi_sox
19 Posts
0
June 16th, 2015 20:00
dondy, Were you able to get this going? Please post your results and if any future assistance is needed.
dondy1
7 Posts
0
June 16th, 2015 20:00
Hi,
Unfortunately No
I don't get the documentation that I need yet 
chi_sox
19 Posts
0
June 16th, 2015 21:00
Let me see if we can get this posted on the community.
dondy1
7 Posts
0
June 16th, 2015 21:00
Hi,
Finally I can get the documentation and start configuring as stated. However, in all my MDMs server I've already run the script but when I tried to shutdown one of the MDM, the service doesn't seems to run. Not even trying to change the cluster to single mode. It just won't do anything.
Anything that I miss maybe?
Regards,
Dondy
chi_sox
19 Posts
0
June 17th, 2015 04:00
Since you tried this before specifying the failover MDMs I would suggest deleting the service, see bottom of page 3, and than attempt to rerun the command. Make sure that you are enter the IPs correctly in the CLI too. If you have a single IP per MDM it should look like this.
[mdm1_ip]+[mdm2_ip]
If you have multiple IPs per it should look like
[mdm1_ip1,mdm1_ip2]+[mdm2_ip1,mdm2_ip2]
dondy1
7 Posts
0
June 17th, 2015 05:00
[root@scaleio-01 bin]# scli --query_cluster
Mode: Cluster, Cluster State: Degraded, Tie-Breaker State: Connected
Primary MDM IP: 10.211.55.30
Secondary MDM IP: 10.211.55.33
Tie-Breaker IP: 10.211.55.32
Management IP: 10.211.55.30, 10.211.55.33
Name: N/A
[root@scaleio-01 bin]#
State Degraded is because I turn off the secondary MDM to simulate crash.
dondy1
7 Posts
0
June 17th, 2015 05:00
Hi,
I did make sure that the previous mdm_failover service are off before I'm putting the new configuration.
Now I can get to the point where the new configuration give me this
Found 3 MDMs' IPs
Found 1 Tie-Breakers' IPs
Successfully started mdm_failover as a service, to stop it use /opt/emc/scaleio/mdm_failover/bin/delete_service.sh
[root@scaleio-04 bin]#
But then again when I simulate to have one MDM (both primary or secondary) goes down, nothing happen.
Any ideas ?
chi_sox
19 Posts
0
June 17th, 2015 05:00
What is the current cluster state?
scli --query_cluster
chi_sox
19 Posts
0
June 20th, 2015 07:00
Just got around to testing this out in my lab and it does work for primary and secondary MDMs as well as the TB.
A few things to check:
- Make the MDM software is installed on the one that will be the standby.
- Check to make sure the service is running on all MDMs
# systemctl list-units --type service --all |grep mdm
mdm.service loaded active running scaleio mdm
mdm_failover.service loaded active running scaleio mdm_failover
mdmonitor.service loaded inactive dead Software RAID monitoring and management
Keep us posted on what you find.