The problem is with Navisphere Agent running on the SP. This is a different program than navicimon, which is the Management Server. In flare releases prior to release 24, the only "safe" way to restart Agent is to reboot the SP. After release 24, Agent became part of Management Server.
You can check on whether or not Agent is running by issuing the navicli command"
navicli -h IP_addess_SPx getagent
x is the IO address for the SP that is showing unmanaged. If you do not receive a reply from Agent, then Agent is not running and will need to be restarted (reboot SP).
If you say you restarted the mgmt server i presume , you have proper network connctvy from mgmt host to the array .
have you tried checking it from some other host .ie putting in IP add of the array's SP in some other host and check if the navisphere is showing the same problem .
I suspect the mgmt host to be a problem becoz the entry of cx600 is coming 2 times in Navisphere and the host cant talk to the agent on the SP .
I would have tried doing a EMC remote on the SP and running the KILL command for the navisphere process on the SP itself
I would have tried doing a EMC remote on the SP and running the KILL command for the navisphere process on the SP itself
for that you have to have a EmcRemote installed and know SP passwords. I don't think EMC takes lightly customers casually connecting to SPs and killing navimom process . I would open a case at this point.
Thank you for the information. I openned a case with EMC support before I took this to the forum. Getting the issues fixed via EMC support is taking a long time. I am going on day three with this problem. I will continue to work with support. Support was connected to the SP via EMC remote, but they did not run the KILL command.
I have the ability to use EMC remote and connect directly to the SP, but I feel some things are best left to support.
I will give points out, if support uses one of the suggestions to fix the problem.
We tried the kill.exe of navicimom.exe that did not fix the issue. We even tried a net stop on the 10Kgovenor that did not fix it either.
It took. a reboot of the SP to clear the problem. Rebooting SPA fixed the duplicate array entires in Navisphere. Heads up, Rebooting the SP caused more problems than it fixed. After the reboot of SPA, all my servers showed fiber login (NO) in connectivity status, plus when I ran powermt display dev=all on the servers they showed all paths to SPA dead.
Does anyone know a command that will tell a Solaris server to re-establish the fiber login to the array? All I could find is reseat the fiber cable on the server or reboot. Does anyone know of something a little cleaner than the two options I metioned above?
Failover software on is Powerpath. Ranging from 3.0.4 to 4.5. I tried # powermt restore dev=all
Looked like the command hung but after almost two hours it returned listing dead paths. I tried stoping and starting naviagent. Naviagent stopped OK, but hung when I tried to start it.
PowerPath and Agent are two different packages, they work together when running but are not neccessary.
use
"powermt check" to remove dead paths
"powermt config" to rescan - will not work for Windows
"powermt restore dev=all" - for Windows
If you lose an SP and your hosts lose paths, they should failover correctly. Check Knowledgebase article emc99467 on PowerLink for the laest failover settings for each OS.
kelleg
4 Operator
•
4.5K Posts
0
June 12th, 2008 12:00
You can check on whether or not Agent is running by issuing the navicli command"
navicli -h IP_addess_SPx getagent
x is the IO address for the SP that is showing unmanaged. If you do not receive a reply from Agent, then Agent is not running and will need to be restarted (reboot SP).
regards,
glen kelley
hemanand
10 Posts
0
June 12th, 2008 00:00
VKR-f4y7L
17 Posts
0
June 12th, 2008 03:00
have you tried checking it from some other host .ie putting in IP add of the array's SP in some other host and check if the navisphere is showing the same problem .
I suspect the mgmt host to be a problem becoz the entry of cx600 is coming 2 times in Navisphere and the host cant talk to the agent on the SP .
I would have tried doing a EMC remote on the SP and running the KILL command for the navisphere process on the SP itself
dynamox
9 Legend
•
20.4K Posts
0
June 12th, 2008 03:00
running the KILL command for the navisphere process
on the SP itself
for that you have to have a EmcRemote installed and know SP passwords. I don't think EMC takes lightly customers casually connecting to SPs and killing navimom process
tim.koopman
73 Posts
0
June 12th, 2008 06:00
Thank you for the information. I openned a case with EMC support before I took this to the forum. Getting the issues fixed via EMC support is taking a long time. I am going on day three with this problem. I will continue to work with support.
Support was connected to the SP via EMC remote, but they did not run the KILL command.
I have the ability to use EMC remote and connect directly to the SP, but I feel some things are best left to support.
I will give points out, if support uses one of the suggestions to fix the problem.
tim.koopman
73 Posts
0
June 12th, 2008 21:00
We tried the kill.exe of navicimom.exe that did not fix the issue. We even tried a net stop on the 10Kgovenor that did not fix it either.
It took. a reboot of the SP to clear the problem. Rebooting SPA fixed the duplicate array entires in Navisphere.
Heads up, Rebooting the SP caused more problems than it fixed.
After the reboot of SPA, all my servers showed fiber login (NO) in connectivity status,
plus when I ran powermt display dev=all on the servers they showed all paths to SPA dead.
Does anyone know a command that will tell a Solaris server to re-establish the fiber login to the array?
All I could find is reseat the fiber cable on the server or reboot.
Does anyone know of something a little cleaner than the two options I metioned above?
kelleg
4 Operator
•
4.5K Posts
0
June 13th, 2008 11:00
glen
tim.koopman
73 Posts
0
June 13th, 2008 15:00
Failover software on is Powerpath. Ranging from 3.0.4 to 4.5.
I tried
# powermt restore dev=all
Looked like the command hung but after almost two hours it returned listing dead paths.
I tried stoping and starting naviagent. Naviagent stopped OK, but hung when I tried to start it.
kelleg
4 Operator
•
4.5K Posts
0
June 13th, 2008 16:00
use
"powermt check" to remove dead paths
"powermt config" to rescan - will not work for Windows
"powermt restore dev=all" - for Windows
If you lose an SP and your hosts lose paths, they should failover correctly. Check Knowledgebase article emc99467 on PowerLink for the laest failover settings for each OS.
glen