This is an OpenPGP/MIME signed message (RFC 4880 and 3156) --wvEvSwxAIdM3VwWQpgacC17A1rTUvCdIv Content-Type: text/plain; charset=utf-8 Content-Transfer-Encoding: quoted-printable Hi, I'm having some trouble installing oVirt 3.6 hosted engine with gluster. First of here are the questions that arose: Do I need separate gluster volumes for the hosted engine and the other VMs? If yes I'm guessing the same optimizations apply to the /engine volume as to the other volume as described in http://www.ovirt.org/Features/Self_Hosted_Engine_Gluster_Support Post installation the storage tab is completely empty though I did create an NFS ISO domain during the hosted-engine --deploy process. Is that normal behavior? Should the engine gluster volume be manually imported ('Import Domain') or is this volume never visible? If I have a separate VM volume I guess I need to create a 'New Domain'. If I do that the over lay window in the UI that allows me to enter the glusterfs connection details clears. I see an animated circle in the middle but nothing happens. There is no entry in the engine log or the in any host log. If I wait long enough the UI terminates the session and I'm at the login again. Last question (for now, I promise): The engine VM is not visible in the UI. Obviously it's running or otherwise there would be no UI. =46rom a variety of sources I get the impression that this shouldn't be but I'm not sure why the engine VM is not showing itself. Sorry for so many maybe obvious questions but since I'm new to oVirt im still stumbling around in the dark for things that do not work according to plan. In case of what should be shown in the UI I'm still not sure what the plan is. To give some background information about my setup: All machines are running up to date versions of CentOS 7.1 64bit. I've three dedicated machines for the gluster storage. The replica 3 gluster volumes are created on top of thinly provisioned volumes. There are three dedicated hosts for virtualization. On the first I'm trying a hosted-engine --deploy. The other two machines are unused since I didn't get everything working as expected. The engine, installed during the deploy process, is installed from a local ISO Centos image. Installation is done manually. I've created two gists: One listing the commands for the gluster volume: https://gist.github.com/tbihawk/4c92dc246a64990942d3#file-ovirt-gluster-s= etup One listing what I've done to install oVirt: https://gist.github.com/tbihawk/adb0a0e7d2a081d63629#file-ovirt-installat= ion-procedure Any help is highly appreciated since I'm kind of stuck. Thanks! Richard --=20 /dev/null --wvEvSwxAIdM3VwWQpgacC17A1rTUvCdIv Content-Type: application/pgp-signature; name="signature.asc" Content-Description: OpenPGP digital signature Content-Disposition: attachment; filename="signature.asc" -----BEGIN PGP SIGNATURE----- Version: GnuPG v2 iEYEARECAAYFAlZJnsQACgkQnGohgOrO9GEPFACgqp0ck/d1uqrgjUGO9y6pc7f2 yxMAn2FMYGF6cIMEQKHLDB1AoKRFMrLM =XA15 -----END PGP SIGNATURE----- --wvEvSwxAIdM3VwWQpgacC17A1rTUvCdIv--
On Mon, Nov 16, 2015 at 10:15 AM, Richard Neuboeck <hawk@tbi.univie.ac.at> wrote:
Hi,
I'm having some trouble installing oVirt 3.6 hosted engine with gluster.
First of here are the questions that arose: Do I need separate gluster volumes for the hosted engine and the other VMs?
Yes, you need two distinct storage domains.
If yes I'm guessing the same optimizations apply to the /engine volume as to the other volume as described in http://www.ovirt.org/Features/Self_Hosted_Engine_Gluster_Support
Absolutely.
Post installation the storage tab is completely empty though I did create an NFS ISO domain during the hosted-engine --deploy process. Is that normal behavior?
Should the engine gluster volume be manually imported ('Import Domain') or is this volume never visible?
It should be automatically imported but we have a know open bug on that: https://bugzilla.redhat.com/1269768
If I have a separate VM volume I guess I need to create a 'New Domain'. If I do that the over lay window in the UI that allows me to enter the glusterfs connection details clears. I see an animated circle in the middle but nothing happens. There is no entry in the engine log or the in any host log. If I wait long enough the UI terminates the session and I'm at the login again.
This one seams a new issue. Can you please attach VDSM from your host?
Last question (for now, I promise): The engine VM is not visible in the UI. Obviously it's running or otherwise there would be no UI. From a variety of sources I get the impression that this shouldn't be but I'm not sure why the engine VM is not showing itself.
This is a side effect of the Hosted Engine Storage Domain Auto import bug: the engine has to import the hosted-engine storage domain to show the engine VM.
Sorry for so many maybe obvious questions but since I'm new to oVirt im still stumbling around in the dark for things that do not work according to plan. In case of what should be shown in the UI I'm still not sure what the plan is.
To give some background information about my setup: All machines are running up to date versions of CentOS 7.1 64bit. I've three dedicated machines for the gluster storage. The replica 3 gluster volumes are created on top of thinly provisioned volumes.
There are three dedicated hosts for virtualization. On the first I'm trying a hosted-engine --deploy. The other two machines are unused since I didn't get everything working as expected.
The engine, installed during the deploy process, is installed from a local ISO Centos image. Installation is done manually.
I've created two gists: One listing the commands for the gluster volume:
https://gist.github.com/tbihawk/4c92dc246a64990942d3#file-ovirt-gluster-setu...
One listing what I've done to install oVirt:
https://gist.github.com/tbihawk/adb0a0e7d2a081d63629#file-ovirt-installation...
Any help is highly appreciated since I'm kind of stuck. Thanks! Richard
-- /dev/null
_______________________________________________ Users mailing list Users@ovirt.org http://lists.ovirt.org/mailman/listinfo/users
This is an OpenPGP/MIME signed message (RFC 4880 and 3156) --nwq2vrg1o9Ac9ItQ6gxqhJFlbQPlTr42s Content-Type: text/plain; charset=utf-8 Content-Transfer-Encoding: quoted-printable Thanks for the answers! On 11/16/2015 05:48 PM, Simone Tiraboschi wrote:
It should be automatically imported but we have a know open bug on that= : https://bugzilla.redhat.com/1269768
Is there anything I can help with? Resetting and reinstalling isn't a problem. So I would gladly test packages or anything else if necessary.
=20 If I have a separate VM volume I guess I need to create a 'New Domain'. If I do that the over lay window in the UI that allows me to enter the glusterfs connection details clears. I see an animated=
circle in the middle but nothing happens. There is no entry in the engine log or the in any host log. If I wait long enough the UI terminates the session and I'm at the login again. =20 =20 This one seams a new issue. Can you please attach VDSM from your host?
This was my fault I guess. I did create two separate storage volumes but thought I could manage them under the same 'storage domain name'. So I gave the same domain name to the newly added vm storage as I did during the deployment of the engine. It seems that adding a new domain that has the same name as an existing doesn't fail with an obvious error. But this is probably related to the failed auto import of the engine storage domain. Giving the engine storage domain a different name than the vm storage domain makes adding an additional gluster storage volume possible without problems. Cheers Richard --=20 /dev/null --nwq2vrg1o9Ac9ItQ6gxqhJFlbQPlTr42s Content-Type: application/pgp-signature; name="signature.asc" Content-Description: OpenPGP digital signature Content-Disposition: attachment; filename="signature.asc" -----BEGIN PGP SIGNATURE----- Version: GnuPG v2 iEYEARECAAYFAlZLLNkACgkQnGohgOrO9GGsiACdEqoxmiYX91ZrPFFyVTcjxybg iPkAnjLOfEzDlJT+U8v8C36wwS7qjD/o =xzz7 -----END PGP SIGNATURE----- --nwq2vrg1o9Ac9ItQ6gxqhJFlbQPlTr42s--
On Tue, Nov 17, 2015 at 2:34 PM, Richard Neuboeck <hawk@tbi.univie.ac.at> wrote:
Thanks for the answers!
On 11/16/2015 05:48 PM, Simone Tiraboschi wrote:
It should be automatically imported but we have a know open bug on that: https://bugzilla.redhat.com/1269768
Is there anything I can help with? Resetting and reinstalling isn't a problem. So I would gladly test packages or anything else if necessary.
If I have a separate VM volume I guess I need to create a 'New Domain'. If I do that the over lay window in the UI that allows me to enter the glusterfs connection details clears. I see an animated circle in the middle but nothing happens. There is no entry in the engine log or the in any host log. If I wait long enough the UI terminates the session and I'm at the login again.
This one seams a new issue. Can you please attach VDSM from your host?
This was my fault I guess. I did create two separate storage volumes but thought I could manage them under the same 'storage domain name'. So I gave the same domain name to the newly added vm storage as I did during the deployment of the engine. It seems that adding a new domain that has the same name as an existing doesn't fail with an obvious error. But this is probably related to the failed auto import of the engine storage domain.
Giving the engine storage domain a different name than the vm storage domain makes adding an additional gluster storage volume possible without problems.
Thanks for the report, adding Roy who is working on that topic.
Cheers Richard
-- /dev/null
participants (2)
-
Richard Neuboeck -
Simone Tiraboschi