No results from SmartState Analysis in oVirt environment


#1

I’ve deployed miq (anand-1.20150308211513_ea4c62a) and oVirt Engine (Version: 3.5.1.1-1.el6). I’m using NFS as the datastore. When I run a smartstate analysis, I receive no errors, however the “last analyzed” field always shows ‘none’ and none of the expected data shows up under the vm properties (no user info, config info, etc). The appliance has all roles checked save for db replication, and has all of the roles active and running, including smart proxy.

I have authenticated correctly against both the hosts and the engine.

What do I need to do to get this working, please?


Error with SSA on RHEV VM on NFS
#2

Have you made sure you also within Infrastructure > Virtual Machine > Select Appliance and then Configure > Edit Edit Management Relationship, to associate this appliance as the one show on the roles page.

Once you do that, also make sure the appliance is able to mount the NFS mount points manually to assure you have the necessary ACL’s on the NFS server to allow for the appliance to mount. Also, make sure the Provider is in the same Zone as the appliance. Lastly, you can check to see if you have SmartProxy Affinity set up to only do certain datastores in the region. Go to Configure > Configuration > Select the Zone, and click on the Smart Proxy Affinity.

You should also log into the appliance and provide a section of the evm.log that shows an error related to performing the SmartState Analysis.


#3

Yes, the management relationship was and is set. I see the appliance has mounted (4x) the remote share. I have only one zone. This is a new installation.

The smart proxy affinity showed nothing checked. I will try that and hope it resolves the issue.


#4

This didn’t resolve the problem. I see these error lines in evm.log:
[----] E, [2015-03-09T17:57:52.318778 #12167:5a080c] ERROR – : MIQ(abstract_adapter) Name: [MiqSmartProxyWorker Load], Message: [PGError: server closed the connection unexpectedly
…: SELECT “miq_workers”.* FROM “miq_workers” WHERE “miq_workers”.“type” IN (‘MiqSmartProxyWorker’) AND “miq_workers”.“id” = $1 LIMIT 1…]
[----] E, [2015-03-09T17:57:52.333451 #12167:5a080c] ERROR – : MIQ(SmartProxyWorker) ID [63336] PID [12167] GUID [329bee5c-c680-11e4-8fd1-001a4af7ce03] Error in update_worker_record_at_exit: Database statement error encountered


#5

Have you checked the status of the scan task?

If you go to configuration->tasks it will show you the current status of all scan tasks (performed in the last 24 hours - by default).

After initiating a new scan, you can go to this page to monitor its progress. The scan may not appear in the list initially, but it should after refreshing a few times. The screen must be refreshed to show any status changes - just make sure you use the refresh button on the application screen instead of the browser’s refresh button.


#6

aha. That’s what I needed. So I see all of my scans failling with errors around paths down the nfs tree.

No such file or directory -
/mnt/vm/4ad24d2a-c687-11e4-8822-001a4af7ce03/rhev/data-center/00000002-0002-0002-0002-00000000003f/mastersd/master/vms/3965bca1-8804-4b6e-b368-0b1a38ba87c6/3965bca1-8804-4b6e-b368-0b1a38ba87c6.ovf
for
VM:[/rhev/data-center/00000002-0002-0002-0002-00000000003f/mastersd/master/vms/3965bca1-8804-4b6e-b368-0b1a38ba87c6/3965bca1-8804-4b6e-b368-0b1a38ba87c6.ovf]

pwd; ls
/mnt/vm/4ad24d2a-c687-11e4-8822-001a4af7ce03/rhev/data-center/00000002-0002-0002-0002-00000000003f/mastersd/master/vms
total 0


#7

So now the question is, why are we looking for files that are not there - or - why are files not there when they should be?


#8

Do you know what kind of backend storage that nfs share resides on? Is it a NetApp filer by any chance?


#9

Looks like the same issue that I am hitting here : Error with SSA on RHEV VM on NFS. My NFS backend is RHEL 7.1 server.


#10

It’s simply NFS exported from an el7 NFS v4 server.


#11

This is a shot from the hip, but did you try the “no_root_squash” option on
the NFSv4 server?


#12

I did. But it does not change anything.


#13

I´m having the exactly the same issue, does anyone have already an idea or workaround?


#14

Just a guess but selinux on either the appliance or nfs server could be rejecting access and manifesting as “no such file or directory”. Maybe try again and check the audit log in /var/log on both afterwards.


#15

Selinux is not in the mix here

Sent using CloudMagic


#16

I think we need help from RHN R&D (ManageIQ & Ovirt), below my 10 cents of an RCA:

ManageIQ tries to access the ovf File on the NFS Share which does not exist on Ovirt 3.5.x at all. I think oVirt changed in 3.5 their behavior by introducing the Disk for “OVF_STORE”, so maybe our issue is related to it…(BTW OVF_STOREs are images that contain the meta-data of VMs and Disks that reside on that data-domain.)

Refresh Relationship:
[----] I, [2015-03-14T14:28:48.114775 #1938:83c00c] INFO – : MIQ(EmsRefresh.save_vms_inventory) EMS: [oVirt], id: [1] Creating Vm [centos-7-cesystems-at] location: [3c2fb30c-16a3-44c3-9526-d8722cb8028a.ovf] storage id: [5] uid_ems: [3c2fb30c-16a3-44c3-9526-d8722cb8028a] ems_ref: [/api/vms/3c2fb30c-16a3-44c3-9526-d8722cb8028a]

Perform Smart State Analyse:
[----] E, [2015-03-14T14:40:20.281073 #1989:c7680c] ERROR – : Q-task_id([edfc1e6c-ca57-11e4-9544-0001a4ad399f]) ScanMetadata error status:[16]: message:[No such file or directory - /mnt/vm/edfc1e6c-ca57-11e4-9544-0001a4ad399f/rhev/data-center/00000002-0002-0002-0002-0000000002d2/mastersd/master/vms/3c2fb30c-16a3-44c3-9526-d8722cb8028a/3c2fb30c-16a3-44c3-9526-d8722cb8028a.ovf]


#17

Any news on this subject since March ?


#18

I’m having the same issue. Ovirt 3.6, manageIQ botvinnik. Under configuration > tasts, it is reported

No such file or directory - /mnt/vm/a1e91102-92f6-11e5-ac15-000c29f24d0c/rhev/data-center/00000001-0001-0001-0001-000000000202/mastersd/master/vms/1f1261b0-7842-48b5-98f6-14ef90c3f1e7/1f1261b0-7842-48b5-98f6-14ef90c3f1e7.ovf for VM:[/rhev/data-center/00000001-0001-0001-0001-000000000202/mastersd/master/vms/1f1261b0-7842-48b5-98f6-14ef90c3f1e7/1f1261b0-7842-48b5-98f6-14ef90c3f1e7.ovf]

Looking at server side, there is only one ovf file - but with different UUID. So it is must bo either for a diffrent oVIRT VM or maybe template? I’m looking directly at xfs mount point, not via NFS.

[root@ovirt-manager ~]# ll /srv/export_domain/780cdc03-9ac1-4aac-bb79-40977ba8fbd4/master/vms/
total 0
drwxr-x---. 2 vdsm kvm 53 nov 24 13:33 2a701f02-be16-40c4-bfd5-5259421aedf0
[root@ovirt-manager ~]# ll /srv/export_domain/780cdc03-9ac1-4aac-bb79-40977ba8fbd4/master/vms/2a701f02-be16-40c4-bfd5-5259421aedf0/
total 8
-rw-r-----. 1 vdsm kvm 4239 nov 24 13:33 2a701f02-be16-40c4-bfd5-5259421aedf0.ovf

Is there a way to get SmartState working?


#19

It works in capablanca rc2, so…