How can I find a host causing a Navisphere Err: Estream construct failed.
Since saturday we have a new err msg on our Clarrion SP that comes up every 20 minutes.
Agent Main exception --Receive request failed. Err: Estream construct failed. Err: [STREAM] Bad input stream Navisphere Agent.
I checked out the emc151541 and it says this can be caused by an earlier version level of the navisphere or host agent. I checked out with the production teams and no changes have been made on the servers this week-end. Is there a way I can find on the Clariion which host is causing this error message.
This could be caused by a network issue - if you look in the Storage tab in Navisphere, open the "Hosts" item below the array serial number (or name for the array) - are all the hosts listed correctly - correct IP addresses and there are no "U" for any of the hosts?
I believe that in the spcollects there is a log that will provide more detail on which host is causing this error.
Do you know if there are any other devices in the network (IP) that is running EMC software - say ECC or a Celerra - some of the other devices have agent programs that can talk to the clariion.
sorry for the delay to respond, but I'm on multitasking and this one is on low priority.
We don't have ECC yet. Our NAS are brand new EMC NS42G with the latest DART.
I ran an spcollect and made a search of the keyword "Estream" on all the unzipped files.
I found naviagent_Oct-17-09_10-20-55.log. This file contains only the Estream error msgs with no info about the host. It also contains indication that the naviagent has been restarted at 10/17/2009 10:20:59. But the errors have started only at 22:47:52 that day.
I also found SPA_navi_getlog.txt. This one contains also the same infos. Same error msgs with no indication of the host and restart of the naviagent on the SP.
I have check the version of all the hostagents on the servers connected to the SAN and all are compatible with Flare 4.28 we are running on the SP. We have no logs of hostagent been installed on another machine not using the SAN.
So no luck.
I will run the keyword search on spcollect of previous months to see if the problem had came before and gone by itself.
Do you have any idea of another keyword that could help on the search? Or other ideas on how to find the host causing the problem.
I would suggest opening a case at this point - attach the spcollects to the case when you open it. We may need to contact engineering about where the message is coming from.
Sorry, I didn't had time to open a support request. The problem message went off by itself after 12 hours and our management didn't want to pursue (pay for) the subject further. So I had no fix from support.
Quick update. I found my problem. The problem was with the agent running on my call home system. Once I restarted the Navisphere agent on the EMC call home server. The Agent Main exception messages stopped.
I do not know want caused the agent to get hung up.
kelleg
6 Operator
•
4537 Posts
1414
0
Posted October 19th, 2009 13:00
glen