This post is more than 5 years old
190 Posts
0
4036
Centera offline - ETA emc250775
I am running 4.1.0-3343-0-0 and am having some problems with the nodes offline. Without getting into many details, support has lapsed and am working on getting that corrected so I can upgrade to a newer version. That said, how would I go about restarting all of the nodes? I know there is a shutdown command but some of the nodes are showing up in the viewer as offline and one cluster is completely unreachable over the network. These are not in production now and won't be until I get the code upgraded but I was wondering if someone had an idea on how to shutdown/start this back up in the mean time.
holgerjakob_c0722c
337 Posts
0
March 3rd, 2011 04:00
Hi Dan
It is not the nice way to power down and power up the system without performing a shutdown in Centera Viewer. We've experienced quite some power problems given that we manage more than 70 systems and so far CentraStar has handled them all.
If it fixes your problem with the systems, toggle the switch in the cabinets.
I suggest that you contact me at holger.jakob@informatio.ch. We'll monitor your systems for a couple of months for free to get you started safely again. You'll need a management station able to send emails with connectivity to the Centera systems. We'll take care of the rest.
Best regards, Holger
holgerjakob_c0722c
337 Posts
1
March 2nd, 2011 03:00
Hi Dan
In Centera Viewer there is a shutdown command. It is under the Commands Menu point. After that you'll have to go in your datacenter to toggle the power switch or the nodes.
Alternatively use the nodelist for the system you can still reach. For the nodes right-click which will display a possibility to restart CentraStar or the Platform OS. Restart the Platfrom OS.
Holger
DanJost
190 Posts
0
March 2nd, 2011 08:00
Thanks for the reply. This works for the ones that I can reach/showing online but the ones showing offline in the viewer I don't have the option to restart. Interestingly enough, though, one of the storage nodes came back online on its own last night. Perhaps the IPMI memory leak referenced in the ETA finally hit the limit and it restarted. With two out of the eight nodes working I can't do much with it and even though the second management node at the other site is "reachable" it won't let me log into it. Unless you think pushing the reset button (referenced in a document on Powerlink which says only to do it under the direction of support) won't cause much trouble (it's already down!) I think I might wait until it is back under support. I don't have a lot of experience with this thing and I don't know how finicky it is.
Dan
DanJost
190 Posts
0
March 7th, 2011 06:00
Thanks for the kind offer - all of the nodes but one came back online on their own. The last one came back after a restart. I think I'll have my support ironed out shortly directly with EMC so after I get it upgraded this particular problem should no longer be an issue.
Dan
holgerjakob_c0722c
337 Posts
0
March 10th, 2011 07:00
Hi Dan
IF you want we can still keep an eye on your systems with our managed services software. We would detect restarts of CentraStar and offline nodes on a daily basis. It would also give you a growth estimation as to how long your capacity and object count last and if your pools are in sych when using replication.
Best regards, Holger