Description of problem: ------------------------ The additional mount options for glusterfs volume was not asked interactively with 'hosted-engine --deploy'. Gluster deployment is done through cockpit. The HostedEngine deployment is guided from Gluster deployment, as the additional config file is passed, which has the following: # cat /tmp/he-answer.conf [environment:default] OVEHOSTED_STORAGE/storageDomainConnection=str:host1:/engine OVEHOSTED_STORAGE/mntOptions=str:backup-volfile-servers=host2:host3 But if the user has made some mistake with the hosted-engine setup, and restarts the 'HE standard deployment' again via cockpit, then there is no way to pass this additional mount options to the hosted-engine --deploy Version-Release number of selected component (if applicable): ------------------------------------------------------------- ovirt-hosted-engine-setup-2.1.0.1-1.el7ev.noarch How reproducible: ----------------- Always Steps to Reproduce: ------------------- 1. Start with Gluster deployment flow. 2. Continue with HE deployment 3. Close the HE deployment in the middle and start over again with 'Standard deployment' Actual results: --------------- additional mount options couldn't be passed Expected results: ----------------- additional mount options should be passed to 'hosted-engine --deploy' Additional info: ----------------- The additional mount options is required for gluster volume, just to add the backup-volfile-servers. If the hosted-engine --deploy has this ability to deduce the backup-volfile-servers ( using 'gluster volume info' or anything related ) and uses it as additional mount options, then this requirement goes void.
Moving the product to 'RHV' as this issue is not seen with upstream ovirt-hosted-engine setup.
(In reply to SATHEESARAN from comment #1) > Moving the product to 'RHV' as this issue is not seen with upstream > ovirt-hosted-engine setup. Apologies for the confusion, the issue exists in upstream oVirt as well. Moving this bug back to oVirt
Considering Hosted-Engine deploy option in Cockpit UI doesn't have a provision to pass answer file, Can we ask the mount options in the interactive question? Note: When HE deployment is triggered from gdeploy wizard in Cockpit, mount options and other volume details passed through answer files. But the same thing is not possible when user 'Restart' the HE deployment in UI.
(In reply to Ramesh N from comment #4) > Considering Hosted-Engine deploy option in Cockpit UI doesn't have a > provision to pass answer file, Can we ask the mount options in the > interactive question? This conflicts with https://bugzilla.redhat.com/1402417 Please consider reopening it.
Don't we add this option ('backup-volfile-servers=host2:host3') to gluster in any case? I remember adding this feature in 3.6.
I see Sahina added it on https://gerrit.ovirt.org/66306, perhaps she can shed more light on the subject
(In reply to Yaniv Dary from comment #6) > Don't we add this option ('backup-volfile-servers=host2:host3') to gluster > in any case? I remember adding this feature in 3.6. We add mount options when user creates glusterfs storage domain from UI/Engine. But for hosted-engine, we requires this option in hosted-engine-setup. Currently mount option can be passed only through answer file in hosted-engine --deploy. But Hosted-Engine deployment in Cockpit doesn't allow user to pass answer files. So we need this option in interactive hosted-engine --deploy script.
(In reply to Ramesh N from comment #8) > (In reply to Yaniv Dary from comment #6) > > Don't we add this option ('backup-volfile-servers=host2:host3') to gluster > > in any case? I remember adding this feature in 3.6. > > We add mount options when user creates glusterfs storage domain from > UI/Engine. But for hosted-engine, we requires this option in > hosted-engine-setup. Currently mount option can be passed only through > answer file in hosted-engine --deploy. But Hosted-Engine deployment in > Cockpit doesn't allow user to pass answer files. So we need this option in > interactive hosted-engine --deploy script. Why do you require it ans why do you not mind losing it in import of the HE SD?
We do not want to lose the mount options while importing HE SD. In fact, these mount options need to be available for additional HE install on other hosts as well. The way I understand it, currently the HE mount options are present in he.conf file on each host, and as per Comment 2 are also stored in HE domain in database. Is this not accurate?
(In reply to Sahina Bose from comment #10) > We do not want to lose the mount options while importing HE SD. In fact, > these mount options need to be available for additional HE install on other > hosts as well. > The way I understand it, currently the HE mount options are present in > he.conf file on each host, and as per Comment 2 are also stored in HE domain > in database. > > Is this not accurate? No, the way it works is that the engine queries the peers for the Gluster volume when adding the SD. Not when you import the HE storage. I gave the test steps on BZ #1434423.
Moving to 4.2 since the bug #1373930 has been targeted to 4.2 and this one depend on it.
(In reply to Sandro Bonazzola from comment #13) > Moving to 4.2 since the bug #1373930 has been targeted to 4.2 and this one > depend on it. Moving to ASSIGNED for the time being.
Everything should be there on hosted-engine-setup side for the vintage flow and for the new one. I'll open a test only BUG on cockpit to ensure that everything is there as well.
Is this a HC specific?
(In reply to Nikolai Sednev from comment #16) > Is this a HC specific? No, it should work also for gluster on external hosts although it's by far more relevant on HC.
[ INFO ] changed: [localhost] Please specify the storage you would like to use (glusterfs, iscsi, fc, nfs)[nfs]: glusterfs Please specify the full shared storage connection path to use (example: host:/path): gluster01.scl.lab.tlv.redhat.com:/nsednev_he_1 If needed, specify additional mount options for the connection to the hosted-engine storagedomain []: backup-volfile-servers=gluster02.scl.lab.tlv.redhat.com:gluster03.scl.lab.tlv.redhat.com [ INFO ] Creating Storage Domain [ INFO ] TASK [Gathering Facts] [ INFO ] ok: [localhost] [ INFO ] TASK [include_tasks] [ INFO ] ok: [localhost] [ INFO ] TASK [Obtain SSO token using username/password credentials] [ INFO ] ok: [localhost] [ INFO ] TASK [Fetch host facts] [ INFO ] ok: [localhost] [ INFO ] TASK [Fetch cluster id] [ INFO ] ok: [localhost] [ INFO ] TASK [Fetch cluster facts] [ INFO ] ok: [localhost] [ INFO ] TASK [Fetch datacenter facts] [ INFO ] ok: [localhost] [ INFO ] TASK [Fetch datacenter id] [ INFO ] ok: [localhost] [ INFO ] TASK [Fetch datacenter_name] [ INFO ] ok: [localhost] [ INFO ] TASK [Add nfs storage domain] [ INFO ] skipping: [localhost] [ INFO ] TASK [Add glusterfs storage domain] ^C Please specify the storage you would like to use (glusterfs, iscsi, fc, nfs)[nfs]: glusterfs Please specify the full shared storage connection path to use (example: host:/path): gluster01.scl.lab.tlv.redhat.com:/nsednev_he_1 If needed, specify additional mount options for the connection to the hosted-engine storagedomain []: backup-volfile-servers=gluster02.scl.lab.tlv.redhat.com:gluster03.scl.lab.tlv.redhat.com [ INFO ] Creating Storage Domain [ INFO ] TASK [Gathering Facts] Deployment was interrupted manually and then I've got again gluster storage option and received mount options again. Works for me forth to provided reproduction steps. ovirt-hosted-engine-ha-2.2.7-1.el7ev.noarch ovirt-hosted-engine-setup-2.2.13-1.el7ev.noarch rhvm-appliance-4.2-20180202.0.el7.noarch Linux 3.10.0-861.el7.x86_64 #1 SMP Wed Mar 14 10:21:01 EDT 2018 x86_64 x86_64 x86_64 GNU/Linux Red Hat Enterprise Linux Server release 7.5 (Maipo) Moving to verified.
This bugzilla is included in oVirt 4.2.2 release, published on March 28th 2018. Since the problem described in this bug report should be resolved in oVirt 4.2.2 release, it has been closed with a resolution of CURRENT RELEASE. If the solution does not work for you, please open a new bug report.