Re: New to oVirt - HE and Metric Store Question
by Strahil Nikolov
I can't recall the exact issue that was reported in the mailing list, but I remember that the user had to power off the engine and the VMs... the Devs can clearly indicate the risks of running the HostedEngine with other VMs on the same storage domain.
Based on Red Hat's RHV documentation the following warning is clearly indicating some of the reasons:
Creating additional data storage domains in the same data center as the self-hosted engine storage domain is highly recommended. If you deploy the self-hosted engine in a data center with only one active data storage domain, and that storage domain is corrupted, you will not be able to add new storage domains or remove the corrupted storage domain; you will have to redeploy the self-hosted engine.
Another one from Red Hat's Self-Hosted Engine Recommendations:
В понеделник, 28 декември 2020 г., 19:40:15 Гринуич+2, Nur Imam Febrianto <nur_imam(a)outlook.com> написа:
What kind of situation is that ? If so, how can I migrate my hosted engine into another storage domain ?
Regards,
Nur Imam Febrianto
From: Strahil Nikolov
Sent: 29 December 2020 0:31
To: oVirt Users; Nur Imam Febrianto
Subject: Re: [ovirt-users] New to oVirt - HE and Metric Store Question
> 1. Right now we are using one SAN with 4 LUN (each mapped into 1 >specific volume) and configure the storage domain for each LUn (1 LUN = 1 >Storage Domain). Is this configuration are good ? One more, about Hosted >Engine, when we setup the cluster, it provision one storage domain, but >the storage domain is not exlusively used by Hosted Engine, we use it too >for other VM. Are this OK or it have a side impact ?
Avoid using HostedEngine's storage domain for other VMs.You might get into situation that you want to avoid.
Best Regards,
Strahil Nikolov
4 years
Cannot run yum, no internet connection on a oVirt node
by jenia.ivlev＠gmail.com
Hello.
I need help setting up the networking on the "self hosted engine" VM.
I can't run yum in the self-hosted engine [1], there seems not to be any internet connection. I get the error "couldn't resolve host: mirrors.fedoraproject.org. Same thing when I do cURL google.com for example. Also I can't ssh to this VM from my physical machine either. I get "connection refused". I can ping the "self-hosted engine" VM though.
Here is the output of ip addr on the "self hosted engine" VM:
2. enp1s0: .... state UP
link/ether 52.54....
3. virbr0: ... state DOWN
inet 192.168.122.1
Here is the ip addr output on the host machine:
3. virbr0: ... state UP
inet 192.168.122.1
Can someone please help me setup networking on the "self hosted engine" VM.
Thanks kindly.
Happy holidays.
[1] https://www.ovirt.org/documentation/installing_ovirt_as_a_self-hosted_eng...
4 years
Re: New to oVirt - HE and Metric Store Question
by Strahil Nikolov
> 1. Right now we are using one SAN with 4 LUN (each mapped into 1 >specific volume) and configure the storage domain for each LUn (1 LUN = 1 >Storage Domain). Is this configuration are good ? One more, about Hosted >Engine, when we setup the cluster, it provision one storage domain, but >the storage domain is not exlusively used by Hosted Engine, we use it too >for other VM. Are this OK or it have a side impact ?
Avoid using HostedEngine's storage domain for other VMs.You might get into situation that you want to avoid.
Best Regards,
Strahil Nikolov
4 years
Moving VMs from 4.3.9 to 4.4.4
by Diggy Mc
What is the best way to move VMs from current 4.3.9 to a new 4.4.4. environment?
The current environment is running oVirt 4.3.9 (started as 4.3.7). There is a single data domain using NFS on a single NAS.
The future oVirt 4.4.4 environment will also have a single data domain using NFS on the same NAS as the current environment.
What is the best way to move approximately 40 VMs from current 4.3.9 to a new 4.4.4. environment?
Are there any caveats or considerations I need to take into account when moving the VMs?
Your help, as always, is greatly appreciated.
4 years
New to oVirt - HE and Metric Store Question
by Nur Imam Febrianto
Hi,
We’re migrating our 10 physical host with 50 small VM to oVirt from VMware and so happy with its performance. There are two question that we want to ask :
1. Right now we are using one SAN with 4 LUN (each mapped into 1 specific volume) and configure the storage domain for each LUn (1 LUN = 1 Storage Domain). Is this configuration are good ? One more, about Hosted Engine, when we setup the cluster, it provision one storage domain, but the storage domain is not exlusively used by Hosted Engine, we use it too for other VM. Are this OK or it have a side impact ?
2. We’re kind of interested using Metric Store (Need performance metric history over the time), is it possible to deploy the Metric Store using smaller resources ? Our hardware is so small, we cant afford to use that lot of resources for the Metric Store.
Thanks before.
4 years
power management iLO5
by ozmen62＠hotmail.com
Hi,
I've red mail list about IPMI fencing. But clould not find a solution.
We have HP gen 10 servers. they have iLO5 and fencing does not work.
In redhat there is a solution
Issue
No iLO5 fencing method available in RHV-manager
Power Management via IPMILAN does not work on HPE Gen 10 systems
Power Management via iLO4 does not work on HPE Gen 10 systems
Resolution
The IPMI interface has to be enabled in the security section of a iLO-5 on a HPE Gen 10 system in order to use the fence-agents iLO3, iLO4 or IPMILAN.
A user with Login permissions and Operator level access for IPMI should be sufficient for fencing.
Root Cause
The default settings of iLO-5 on a HPE Gen 10 system have disabled IPMI access to the management controller. The fence agents for iLO-3 onward default to be using IPMI in order to communicate with the management controller. Therefore iLO3, iLO4 as well as IPMILAN fail to connect to a iLO-5 management controller.
But, there is no IPMI option in power management.
How can i enable power managemenet properly with iLO5 or IPMI?
Thanks
4 years
cinder backup fail
by ozmen62＠hotmail.com
Hi,
After 4.4.4 upgrade , i've enabled grafana.
Now, there is an error while backing up
Backing up:
Notifying engine
- Files
- DWH database 'ovirt_engine_history_20201223073120'
- CINDERLIB database 'ovirt_cinderlib'
Notifying engine
FATAL: Database ovirt_cinderlib backup failed
in log file, it says;
pg_dump: error: connection to database "ovirt_cinderlib" failed: FATAL: Ident authentication failed for user "ovirt_cinderlib"
could you help to fix it?
4 years
4.4.4 hosted engine deployment failure
by Diggy Mc
While trying to deploy a 4.4.4 hosted engine via the cockpit interface, I received the following error(s):
[ INFO ] TASK [ovirt.ovirt.hosted_engine_setup : Obtain SSO token using username/password credentials]
[ INFO ] ok: [localhost]
[ INFO ] TASK [ovirt.ovirt.hosted_engine_setup : Wait for the host to be up]
[ INFO ] TASK [ovirt.ovirt.hosted_engine_setup : Notify the user about a failure]
[ ERROR ] fatal: [localhost]: FAILED! => {"changed": false, "msg": "Host is not up, please check logs, perhaps also on the engine machine"}
...
[ INFO ] TASK [ovirt.ovirt.hosted_engine_setup : Notify the user about a failure]
[ ERROR ] fatal: [localhost]: FAILED! => {"changed": false, "msg": "The system may not be provisioned according to the playbook results: please check the logs for the issue, fix accordingly or re-deploy from scratch.\n"}
I am trying to build a new oVirt 4.4.4 environment using Oracle Linux 8.3 host servers. I created /var on a separate LV from / (root). I also configured adapter bond per the docs and the same way I did with my 4.3.7 environment.
Did I miss a configuration requirement for the OL 8.3 installation? Where can I find more details about the error(s)?
4 years
