This post is more than 5 years old
3 Posts
0
1069
June 20th, 2012 02:00
Navisphere GUI shows only one host even there are two servers...
Hi All,
I would like to get help from you about some issue that I am dealing with at the moment. I was doing fresh installation on some SUN servers. Let me write the steps that I followed and please find my question about EMC at the end of them:
1) I have Installed new Sun Solaris OS and the softwares which is required for EMC into first Server.Lets assume that; Ip adress is 10.0.0.1 and hostname is A for this server.
2)After completed all installation and was able to manage the LUNs on this server, I decided to change IP adress and hostname of this server to another second Server(assume that second Server hostname is B , IP address is 10.0.0.2). But please be informed that server B has already connected to the EMC.
3) I have swapped the IPs and hostnames between Server A and B. But since I am not expert I have not modified agentID.txt file on those servers.
4) Now, volume mount points seems ok and working properly on Servers , I can reach all the files but there is something wrong on the Navisphere GUI . I can see that those servers have the same host and under the "Hosts" subtree, it shows me that "A[10.0.0.1;Fibre;Host Agent not reachable]". However even server A and server B has different storage group on EMC, there is only one and same host under "Hosts" tree.
So I need to correct this issue. I performed some search on the net and EMC forums and found that I need to modify agentID.txt with the correct values on each servers that are effected and restart the" host agent", but I am not sure how I can restart the host agent on Sun servers.However, I do not know, by performing this action may cause data interruption between EMC and Server. If you can help me on this I would really appreciated.
Note: When I do check process, I got below output.I have changed the agentID.txt with the correct values and killed and restart the below process that has "21457" id but it did not modify the "/etc/log/HostIdFile.txt" and so not fix my problem. It is important not to reboot servers as a requirement to fix this problem.
emm:/# ps -ef |grep -i agent
root 21470 10599 0 11:51:51 console 0:00 grep -i agent
root 2235 1 0 14:23:26 ? 1:00 /opt/VRTSobc/pal33/bin/vxpal -a StorageAgent -x
root 21457 1 0 11:51:47 ? 0:00 /opt/Navisphere/bin/naviagent -f /etc/Navisphere/agent.config
Thanks in advance,
melihcelik
3 Posts
0
June 20th, 2012 03:00
Hi All. Seems that I found the solution:
1) I have deleted the HostIdFile.txt file under /etc/log directory.
2) I have corrected agentID.txt file that is located under root directory with the correct values
3) Restarted agent by running below commands:
/etc/init.d/agent stop
/etc/init.d/agent start
4) Checked the GUI and now I can see all the hosts with the correct names and IP adresses.
Only one small issue that I observerd is there is still older non-reachable hostnames is remained in the Hosts list...However it will not effect anything on the current configs and I just leave it like that.
melihcelik
3 Posts
0
June 20th, 2012 04:00
AnkitM has provided some inputs in my discussion area but while I was writing to AnkitM and submit my post I recognized that somehow it has been deleted . So even AnkitM's post does not seems here, I would like to share my post.
Thank you for the link(https://community.emc.com/message/640528#640528#640528) and your answer. If you mean right click on host and "update now" option, after updated the agentID.text file , it did not fix the issue.When I did this ,it was still complaining not to connect host.
Secondly, I will follow the steps you provided to remove the Ghost hostname. However , If I restart SP and If some LUNs are manually managed by either SP-A or SP-B , this action may cause some data transfer failure between server and EMC , what do you think? I am not sure how some LUNs for some other servers are configured earlier. Some of them should only be managed by only one SP, is it possible that someone configured them like that? I am not sure because I have not enough knowledge of EMC. So if it there is some risk, I prefer to not touch ghost host since it should not effect anything.
One more thing I would like to add: In your action plan for Unix, there is some difference when I compare mine. I realized that when I delete the HostIdfile.txt file and stop and start agent; new HostIdfile.txt is created by using the new values on agentID.text. So the difference, we do not need to re-boot the servers. It is really major point for me to not reboot the server in my case.
Really apreciated with your useful inputs.
AnkitMehta
1.4K Posts
0
June 21st, 2012 11:00
Oh! Sorry about that, I myself had deleted that post considering I double posted that. By the time I realised it was too late.
Anyhow, coming back to your query:
Before you reboot any SP's
Check if host is using multipathing software let it be native, MPIO or EMC Powerpath for path management
Secondly, make sure failover settings are correct.
To understand and know about failover see:
https://community.emc.com/docs/DOC-8769
https://community.emc.com/docs/DOC-12817
Moreover, EMC Knowledgebase solution emc99467 details in a single document the Initiator, Arraycommpath and Failovermode settings for PowerPath, DMP, PLLinks, and native failover software attached hosts.
As you mentioned you are new to EMC Storage and not sure about this, I would suggest to open a chat or service request and seek help from EMC Technical Support.
I will get back to you soon about the second query.