swalker_LST Tracker

Re: vSphere 6.0U3 EAM Error

swalker_LST — Thu, 29 Mar 2018 13:27:26 GMT

This was really helpful for me as I just ran into the same problem which was also blocking my 6.5 upgrade. Through discussion with VMware - in addition, line 46 in your example should be updated to include your specific hostID obtained using:

cat /etc/vmware/install-defaults/sca.hostid

After that I found that an additional line was required before the service would start correctly:

eam.web.root=/usr/lib/vmware-eam/web

Finally, in case this was done using the root user you might find it necessary to set the permissions correctly for the properties file if it was deleted/recreated:

chown eam:cis eam.properties
chmod -R 644 eam.properties

Hope this helps if someone else runs into this issue during an upgrade..

Re: No items in datacenter

swalker_LST — Mon, 26 May 2014 14:04:00 GMT

Hey All,

I am happy to update my own part of this discussion in the hope that it might help someone else in the same situation. Basically we were suffering Inventory differences between the vSphere client (C#) and the Web Client, and this now explains why it was not possible for vCOPs to reliably pull information out of vCenter for all of the datacenter nodes/hierarchy as detailed above.

However, since the Inventory service was running and everything else seemed OK we believed that this was a vCOPs problem. Almost by chance, we attempted to restart the vCenter Server and found that the Inventory Service would no longer start up afterwards. This led us to investigate why 40GB of hard disk space on the vCenter Server had been eaten up by Inventory Service DB XHIVE transaction logs. We recreated the Inventory Service DB using the documented procedure, but this did not resolve the problem. We could see more than 10MB of data being written to the Inventory logs per second, and high CPU utilisation. Further investigation of the debug logs for the Inventory Service led us to discover a known problem with the use of the vFlash feature on (only a single server) one of vCenter member hosts was causing the problem. We removed it from vCenter, applied the patch as detailed in the link below, and the problem with vCOPs is now resolved.

I hope this description makes someone else's day a little bit easier!!..

VMware KB: VMware vCenter Server consumes most of the CPU, memory, or disk I/O after enabling Virtual Flash

Re: No items in datacenter

swalker_LST — Tue, 20 May 2014 14:52:29 GMT

Hi again, I removed the original appliances (vApp) by disassociating them from vCentre and removing the asset from the Solutions tab. This was completed successfully before deploying a new vCOPs vApp, and this was again registered with vCentre. So there were never two sets of vCOPs running at the same time.

When I'm referring to a lack of data within vCOPs for a certain datacentre (i.e. nothing at all) what I meant was that when I try to expand the DC name using the drill down option on the left pane I am able to expand it as far as displaying individual cluster names, but no health information is being received from any objects within it. All of the folders within the datacentre are detected, but displayed only as grey squares on the health map.

This was multi-datacentre hierarchy was working fine like this for at least 18 months, and the missing information is consistent with what I couldn't view on the original appliance and the one which I've just deployed in place of it. The following screen capture shows the hierarchy expanded to its full extent, but I can definitely confirm that the permissions are identical for each datacentre.

The only other slightly new thing that we've done recently is set up VM Storage Profiles for use with Storage DRS.

Re: No items in datacenter

swalker_LST — Tue, 20 May 2014 14:33:48 GMT

I'm not very familiar with CustomUI - but since I've redeployed the appliances from clean the only thing that remains in common is the vCenter DB. None of the hosts are marked as maintenance mode in the vSphere Web Client/Client and if I go into CustomUI and choose, Environment>Environment Overview I can see all of the objects within the Inventory (all marked with a 100 health score). All have got a green mark against the Collection column, and the Data receiving icon. When I tried this earlier today using the original appliance almost all of the items were marked with a blue question mark. I would be tempted to leave it and see what happens next, but we have already started receiving data for a smaller subordinate datacentre and nothing at all for any datacentre with a cluster or larger configuration present.

Update: Actually, it's a very long list of items in the CustomUI view - and all of the entries which are not showing up in the vCOPs are marked with a blue question mark.

Message was edited by: swalker_LST

Re: No items in datacenter

swalker_LST — Tue, 20 May 2014 14:18:16 GMT

Over the last couple of days we have also started to experience the same issue. I deployed 5.6 a year or so ago, then upgraded through 5.7 and later 5.8.1 without any issues. We haven't changed any configuration settings since then for either the collector user or the vCOPs license, however recently (..and it may be unconnected) we decommissioned several datastores and added some new ones. Since around that time we started seeing a problem where some datacentres cannot be drilled down into, whereas others can. Our vCenter permissions are applied right at the top of the hierarchy (on the VC object itself) and the propagate option applies exactly the same set of permissions to each of the datacentre objects. Some standalone hosts in the root of the hierarchy can be seen, whereas others cannot.

We have approximately 80% of the inventory items missing now. I unregistered and reregistered the vCOPs solution with vCenter which was successful, but didn't fix the problem. Further to this I restored a backup of the original appliance that was around a month old, and this also has the same issue. There would appear to be some vCenter inventory related issue going on which might have occurred when we added new resources recently, however all of the vCenter health checks work out fine so I'm at a loss as to what has caused it.

Finally, in order to prove whether the issue is VC related I've just deployed an entirely fresh set of vCOPS UI and Analytics VMs - then registered/licensed correctly and confirm exactly the same behaviour is being experienced.

Does anyone have any ideas about where to look next?

Re: vFlash performance metrics missing

swalker_LST — Thu, 13 Feb 2014 17:04:20 GMT

Me too - I've just upgraded from a fresh installation of vCenter Server 5.1 on Oracle DB, to 5.5b and the counters are missing. I have enabled vFRC on one ESXi 5.5 host (Ent Plus license) and I can use the command line to verify cache information but the counters are not visible in either the C# or Web Clients. Don't seem to find any other information from other people having this issue.