On 10/5/18 8:10 AM, Vincent Royer wrote:
Alright, some combination of messing around got me mostly back up and
running here. Iscsi is working reliably, VMs are running on the new
iscsi domain. the old NFS domain is empty save a couple of OVF_STORE
disks.
good job
The procedure now, as I understand it, is to shut down all vms,
why ?
put all storage domains into maintenance
which additional domain you have ?
except the old NFS and new SCSI domain, then push the NFS one into
maintenance, which *should* promote the ISCSI domain to master.
it's going to happen. but it's not importend for ovirt where the master
lives.
After that, I'll try to move the HE storage to iscsi too, although
it would be very helpful to have access to this document
<https://access.redhat.com/solutions/2998291> for that procedure.
Sorry i have also no access, but it sounds like backup/restore ...
RHEV Documentation is <a
href="https://access.redhat.com/documentation/en-us/red_hat_virtualization/4.2/";>here</a>.
I think it's like E(BareMetal) <a
href="https://access.redhat.com/documentation/en-us/red_hat_virtualization/4.2/html/self-hosted_engine_guide/chap-migrating_from_bare_metal_to_a_rhel-based_self-hosted_environment";>(see
here)</a>to HE(ISCSI), but you want HE(NFS) to HE(iSCSI).
It's time for a new mailing thread.
cu
Oliver
thanks for your assistance!
On Thu, Oct 4, 2018 at 2:23 PM Oliver Riesener
<[email protected] <mailto:[email protected]>>
wrote:
Hi Vincent,
nice to hear the news :-)
I have read the BZ and see you run into NFS trouble and solved it now.
I took a look on my centos server for nfs data domains and
see server running V4 and the clients (node mounts with protocol
vers=4.1)
I run the latest (and greatest) ovirt stable 4.2.6.4-1 on centos
7.5+ with
engine installed and a ovirt-node 4.2.6.4.
* If you can migrate your running VMs and can switch your SPM,
i would upgrade and reboot the hosts one by one, now.
* reboot seem to be a minimum, remember you do that `virt. thing´,
therefor you can access and boot your bare metal and host os ;-)
ok, back to iSCSI, i have also a EQUALOGIC running as iSCSI target
over years.
* I have allowed multi host access to the volumes which ovirt uses.
Access control lists contains raw IP addresses from my ovirt-hosts.
ovirt handles the volume access virtuos with multipathd and lvm
vg’s and lv’s.
unused lvs are offline (host specific) and released volumes are
deactivated.
* Also it’s possible you have to reinstall (from GUI) your hosts,
to upgrade or install the needed packages, which handles iSCSI
Client access.
* If you then free from errors and your iscsi data domain still
missing, we talk
about vg activation and domain import.
Sheers
Oliver
Am 04.10.2018 um 22:00 schrieb Vincent Royer
<[email protected] <mailto:[email protected]>>:
Ok, getting somewhere here.
did a rpcinfo -p and found no nfs entries in portmap.
systemctl stop nfs
systemctl start nfs
Suddenly shares are mounted and datacenter is up again.
was able to add export domain over NFS.
Why would nfs shit the bed?
still can't seem to get iscsi mounted properly now, and that's
where all the disks are located :/
On Thu, Oct 4, 2018 at 11:00 AM Vincent Royer
<[email protected] <mailto:[email protected]>> wrote:
Thanks for your help Oliver,
To give you some background here:
Host 1 on Ovirt 4.2 attached to NFS storage
Host 2 I upgraded to Ovirt 4.2.5 and then 4.2.6, since then
it has had troubles with NFS due to this bug
https://bugzilla.redhat.com/show_bug.cgi?id=1595549. The
host was up and could run the hosted engine, but could not
migrate any VMs to it.
I decided to switch from NFS to ISCSI so that I could stay on
current releases. So I began the work of attaching iscsi
domain.
The iscsi domain attached, and I transferred most of the
disks to it. Then it started melting down saying that Host 1
could not mount it, and the whole DC went down.
Current status is data center "non responsive". Keeps trying
"Reconstructing master domain on Data Center" over and over
again but always fails. Master domain status is "inactive".
Clicking activate fails. The new ISCSI domain, I put in
maintenance until I figure the rest out. I can't add or
remove any other domains, Ovirt says I need to attach the
master first.
Both hosts are "UP". Host 1 health is "bad" Host 2 health
is "ok", and it is running HE. Host 1 (the 4.2 host) says
"this host needs to be reinstalled". But the reinstall
option is grayed out.
I am weary about updating host1, because of the NFS storage
bug... I fear it won't ever be able to attach the old domain
again.
If I try mounting the NFS shares in cockpit from either node,
they say "mount.nfs: Remote I/O error". However on another
blank centos machine sitting on the same network, I can mount
the shares normally.
*Vincent Royer*
*778-825-1057*
*
*
<http://www.epicenergy.ca/>
*SUSTAINABLE MOBILE ENERGY SOLUTIONS*
On Thu, Oct 4, 2018 at 1:04 AM Oliver Riesener
<[email protected]
<mailto:[email protected]>> wrote:
When your hosts are up and running and your Domain didn't
go active within minutes
* Activate your Storage Domain under:
Storage -> Storage Domain -> (Open your Domain) -> Data
Center -> (Right Click Your Data Center Name) -> Activate.
On 10/4/18 9:50 AM, Oliver Riesener wrote:
Hi Vincent,
OK you master domain, isn't avail a the moment, but no
panic.
First off all we need the status from your hosts. No
HOSTS -> No Storage !
* Do you reboot them hard, without Confirm "Host has
been rebooted"
* Are they actived in the DataCenter / Cluster ? Green
Arrow ?
On 10/4/18 7:46 AM, Vincent Royer wrote:
I was attempting to migrate from nfs to iscsi storage
domains. I have reached a state where I can no longer
activate the old master storage domain, and thus no
others will activate either.
I'm ready to give up on the installation and just move
to an HCI deployment instead. Wipe all the hosts clean
and start again.
My plan was to create and use an export domain, then
wipe the nodes and set them up HCI where I could
re-import. But without being able to activate a master
domain, I can't create the export domain.
I'm not sure why it can't find the master anymore, as
nothing has happened to the NFS storage, but the error
in vdsm says it just can't find it:
StoragePoolMasterNotFound: Cannot find master domain:
u'spUUID=5a77bed1-0238-030c-0122-0000000003b3,
msdUUID=d3165759-07c2-46ae-b7b8-b6226a929d68'
2018-10-03 22:40:33,751-0700 INFO (jsonrpc/3)
[storage.TaskManager.Task]
(Task='83f33db5-90f3-4064-87df-0512ab9b6378') aborting:
Task is aborted: "Cannot find master domain:
u'spUUID=5a77bed1-0238-030c-0122-0000000003b3,
msdUUID=d3165759-07c2-46ae-b7b8-b6226a929d68'" - code
304 (task:1181)
2018-10-03 22:40:33,751-0700 ERROR (jsonrpc/3)
[storage.Dispatcher] FINISH connectStoragePool
error=Cannot find master domain:
u'spUUID=5a77bed1-0238-030c-0122-0000000003b3,
msdUUID=d3165759-07c2-46ae-b7b8-b6226a929d68'
(dispatcher:82)
2018-10-03 22:40:33,751-0700 INFO (jsonrpc/3)
[jsonrpc.JsonRpcServer] RPC call StoragePool.connect
failed (error 304) in 0.17 seconds (__init__:573)
2018-10-03 22:40:34,200-0700 INFO (jsonrpc/1)
[api.host] START getStats()
from=::ffff:172.16.100.13,39028 (api:46)
When I look in cockpit on the hosts, the storage domain
is mounted and seems fine.
_______________________________________________
Users mailing list [email protected] <mailto:[email protected]>
To unsubscribe send an email [email protected]
<mailto:[email protected]>
Privacy Statement:https://www.ovirt.org/site/privacy-policy/
oVirt Code of
Conduct:https://www.ovirt.org/community/about/community-guidelines/
List
Archives:https://lists.ovirt.org/archives/list/[email protected]/message/LTZ6SIFYDFEMSZ4ACUNVC5KETWG7BBIZ/
--
Mit freundlichem Gruß
Oliver Riesener
--
Hochschule Bremen
Elektrotechnik und Informatik
Oliver Riesener
Neustadtswall 30
D-28199 Bremen
Tel: 0421 5905-2405, Fax: -2400
e-mail:[email protected]
<mailto:e-mail:[email protected]>
Tel: 0421 5905-2405, Fax: -2400
e-mail:[email protected]
<mailto:e-mail:[email protected]>
_______________________________________________
Users mailing list -- [email protected]
<mailto:[email protected]>
To unsubscribe send an email to [email protected]
<mailto:[email protected]>
Privacy Statement: https://www.ovirt.org/site/privacy-policy/
oVirt Code of Conduct:
https://www.ovirt.org/community/about/community-guidelines/
List Archives:
https://lists.ovirt.org/archives/list/[email protected]/message/V72KMULZJAT3XIR3GBTOCA5RLACVQSRC/
_______________________________________________
Users mailing list -- [email protected] <mailto:[email protected]>
To unsubscribe send an email to [email protected]
<mailto:[email protected]>
Privacy Statement: https://www.ovirt.org/site/privacy-policy/
oVirt Code of Conduct:
https://www.ovirt.org/community/about/community-guidelines/
List Archives:
https://lists.ovirt.org/archives/list/[email protected]/message/OZ3TOO554D4YAZJ6FPGI4SAJ6CKWZRFH/
_______________________________________________
Users mailing list -- [email protected]
To unsubscribe send an email to [email protected]
Privacy Statement: https://www.ovirt.org/site/privacy-policy/
oVirt Code of Conduct:
https://www.ovirt.org/community/about/community-guidelines/
List Archives:
https://lists.ovirt.org/archives/list/[email protected]/message/XBVJAWRSA2NCNAWEOFKCONJBP7NTPSF2/
--
Mit freundlichem Gruß
Oliver Riesener
--
Hochschule Bremen
Elektrotechnik und Informatik
Oliver Riesener
Neustadtswall 30
D-28199 Bremen
Tel: 0421 5905-2405, Fax: -2400
e-mail:[email protected]
_______________________________________________
Users mailing list -- [email protected]
To unsubscribe send an email to [email protected]
Privacy Statement: https://www.ovirt.org/site/privacy-policy/
oVirt Code of Conduct:
https://www.ovirt.org/community/about/community-guidelines/
List Archives:
https://lists.ovirt.org/archives/list/[email protected]/message/MRBMBHEXZXLWGLAE7GGA55AL5233ASXC/
