Unsolved
This post is more than 5 years old
19 Posts
0
4216
Host cannot be seen in "Connection Status" in Navisphere
This is the rundown of the issue to where I am not able to see the host in the "connection status" of the CLARiiON. A few things should be noted starting out.
1. I can see the WWN for both fabric adaptors in VCenter.
2. I also see the WWN in the Brocade DCX fabric router. (this is telling me that the HBA is indeed logging in)
The troubleshoot thought process so far has led me to perform the following:
Update all IBM firmware off of the latest boot image loader from IBM's website. Sent AMM logs to IBM for evaluation and they confirmed that everything seems to be fine. All research this far has pointed to mainly two posibilities; impropper zone configuration or firmware on HBA. The zoning I'm positive is correct. After installing the new HBA the emulex device shows completely different WWN so it's easy to see from the other QLogic ones. That WWN is zoned with the same aliases as every other device that is working. To address the HBA issue, I have replaced the QLogic HBA entirely and installed an Emulex HBA and flashed it with the lastest driver built for ESXi 4.0. After that I rescanned both fabric devices in VCenter for Blade10 and Navishere still cannot see the WWN in "connection status" to register the device.
At this point it seems that it's not hardware or firmware related since the issue did not follow after hardware was replaced. In regards to VMWare, I don't see anything in there other than rescanning the physical fabric devices that could be causing this.
Please let me know if anyone sees something I may be missing here.
Rashid
kelleg
4.5K Posts
1
June 23rd, 2010 14:00
Do you see other hosts on the clariion using the same SP ports? If so, it's a Zoning issue - did you save and apply? Do you have sufficient licenses on the switch.
glen
bernd3
42 Posts
1
June 24th, 2010 03:00
When checking the zoning I recommend using the command nszonemember . This directly queries the fabric nameserver and shows all active connected ports.
What firmware are you using on the Switch? There is a bug with state change notification which could cause this behaviour. It was fixed recently in 6.3.1a.
When scanning for luns on the VMware host please follow the steps mentioned in http://communities.vmware.com/thread/8843
hope that helps
bernd
Message was edited by: bernd (typos)
Rashid-GDIT
19 Posts
0
June 24th, 2010 06:00
Also, I read the article in regards to the rescanning of HBA's in VCenter. But I beleive that's addressing VCenter seeing allocated LUNS in which I haven't gotten this far yet. At this point it seems that the issue can't be at the VMWare side because I haven't been able to present any LUNs to the ESXi host yet. VCenter just sees what I present to it really.
dynamox
1 Rookie
1 Rookie
•
20.4K Posts
1
June 24th, 2010 06:00
zone config was saved and activated ?
Rashid-GDIT
19 Posts
0
June 24th, 2010 06:00
See, I felt the same way. That it had to be zoning. But again, I know that the WWN's are correct. Especially after changing out the HBA to the Emulex because like I said it stands out from the other QLogics. This blade Alias is zoned with the exact same aliases for SPA and SPB that the 4 other blades in this same chasis are pointing too in other zones. I have saved the changes and also the zone config several times wondering if it just didn't take. I even rebooting the fabric switches. The particular chasis that these blades are in are all going through the same two SSM modules which I have setup for "Access Gateway" mode which puts it just a passthrough switch. I do see all other HBA's for all other blades in both the switch and in Navisphere. It's just this one that I can see in switch but nothing in Navisphere. I will double check the version of these switches as well as the licensing. Should be ok but at this point it's worth a quick look.
RobertDudley
448 Posts
0
June 24th, 2010 06:00
i have experienced the issue where zoning is correct but the host is just not logging into the storage frame. it seems that the naem server is not being updated and a way to "force" and update is to block/unblock an ISL. I beleive that at the highest firmware levels this issue is "fixed" well mostly anyway. I think someone posted something to this effect.
Rashid-GDIT
19 Posts
0
June 24th, 2010 06:00
Yep, more than once. I save the Active Config and also selected Enable just in case. It is the only Zone config that I have.
dynamox
1 Rookie
1 Rookie
•
20.4K Posts
0
June 24th, 2010 07:00
while he/she is at it ..have him/her restart management server on each SP (non-disruptive ..go to http://SP/setup), maybe Navi is being flaky.
Rashid-GDIT
19 Posts
0
June 24th, 2010 07:00
I'm in a VMWare course now so I'm having a collegue check on that firmware version. I will post the results.
aloski
38 Posts
0
June 24th, 2010 23:00
also make sure you are not using java 6 update 20 as I noticed this update causes navisphere not to show all objects.
bernd3
42 Posts
0
June 25th, 2010 00:00
Did a quick check on the firmware bug mentioned by RobertDudley and me:
This is from the v6.3.1a release notes. I'm mentioning this 'cause this bug gave us a hard time, too with hosts unable to find any luns.
bernd
Rashid-GDIT
19 Posts
0
June 25th, 2010 06:00
Man, I am so eager to get back over there and fix this issue. My collegue hasn't gotten the opportunity yet to check on this issue while I'm here in class. Today is my last day and I will be able to continue and update all on Monday when I return to the lab. Thanks to all for following up on this issue and taking the time to research and post.
kelleg
4.5K Posts
0
June 25th, 2010 09:00
Try the JAVA fix first - that's the quickest - go to 6 update 12/13 or 16
glen
Rashid-GDIT
19 Posts
0
June 28th, 2010 07:00
Ok guys,
I verified that I do have the latest Fabos version 6.3.1a firmware on the Brocade DCX switch. Also, i was running Java version 6.12 but I went ahead and updated it to 6.16 just to see if it would help. So far I still cannot see this blade device in Navisphere. I'm totally out of ideas here. What would be the next logical step in troubleshooting?
Thanks for all your help so far.
Rashid
dynamox
1 Rookie
1 Rookie
•
20.4K Posts
0
June 28th, 2010 13:00
did you restart management server on each SP ?