CloudStack 4.22 – VM deployment on LINSTOR primary fails during ROOT volume population (qemu-img convert), while volume creation succeeds #12388
|
sos_2026-01-02_15-03-28.tar.gz ###Architecture ┌──────────────────────────────────┐ ┌──────────────────────────────────┐ ┌──────────────────────────────────┐ ======================================================================
|
Answered by
bilalinamdar
Jan 23, 2026
Replies: 20 comments 7 replies
|
Thanks for opening your first issue here! Be sure to follow the issue template! |
0 replies
|
I am uploading the logs here. for the kvm's and the mgmt tar cvpzf ~/cskvm0*.logs.tar.gz /var/log/linstor-satellite/ /var/log/linstor-controller/ /var/log/cloudstack/ /var/log/libvirt/ /var/log/syslog /var/log/dmesg /var/log/kern.log tar cvpzf ~/csmgmt01.logs.tar.gz /var/log/cloudstack/ /var/log/syslog /var/log/dmesg /var/log/kern.log cskvm03.logs.tar.gz Also some settings of linstor root@cskvm01:~# linstor node list linstor sp l root@cskvm01:~# linstor sp l Resource Groupslinstor rg l root@cskvm01:~# linstor rg l root@cskvm01:~# On allno other parameter is set for linstor drbd. #Current Issue
But as a additional disk i can use linstor for existing vm created via nfs. it also migrated to another host without isssue live. can also create standalone volumes with linstor. I only can't create VM's with linstor before this on plain install and linstor as a only primary storage even system vm's where not making up. so this time i used nfs as primary than second primary linstor. VM spinup issue. i think this is mostly due to some template clone / raw / qcow2 issue as per some observation in logs. Update: I tried to deploy vm with ALPINE ISO and linstor and it worked. But it is not working with the default template from the cloudstack wizards (qcow2/img) 
root@cskvm01:/var/log/cloudstack/agent# linstor v l |
0 replies
|
In the Linbit forums I suspected that the wrong format in the Database might be a problem for the templates, So I simply suspect, that the templates images themselves are not working? Can you try with some other templates, like simple ubuntu cloud images? |
0 replies
|
Hi @rp- ,
To resolve Point 2 that is deployment of template or Instance making issue i tried to convert the qcow2 image to raw didn't work than i made changes in the database of template to recognize it as raw but still failed. I tried vice versa still failed. In the linstor forum i was blocked to reply but i think it is now operational. That's why i was unable to reply there. If you haven't checked the log yet u can. I would suggest to check FIRST failures not last as u will find why i needed to try to fix it.. https://drive.google.com/file/d/1-RLNwX0lBIACjoB2luSxTncBFj3JHCtk/view?usp=drive_link Logs for the csmgmt01, cskvm01,cskvm02,cskvm03 #MGMT Currently i have installed HCI and checking it. Its working fine out of the box. Will again erase and install the same architecture mentioned here this weekend. And this time i will keep it as is. I mean will not touch much will report out of the box experience. |
0 replies
|
@bilalinamdar And then the first error is already this: But now as I look at it, I found it more interessting that there is already an error mounting the nfs secondary share and I'm a bit surprised that it then even tried to |
0 replies
|
@rp- Thanks for responding. |
0 replies
|
I'm also sharing here my "playbook" from Milan. |
0 replies
|
Hi, This is initial Investigation. So i used virter and replicated exact env as yours from the playbook.
So i investigated further and renamed the .img of ubuntu 24.04 to qcow2 and reuploaded it as new template and its deployment worked. later i also used. Let me check further as this was virter env. but atleast i am happy for now. Will update soon as this is initial investigation 
remember the ubuntu 22 and 24 are from the official wizard download and not manual (.img) To summarize
extra command needed in the poc doc #On main I have seen similar .img issue in proxmox forum somewhere. |
1 reply
|
AFIAK, the ubuntu .img are qcow2 images just "named" bad. So what you are saying is that, the picked ubuntu images from the install wizard are not working with Linstor storage? I would guess than that it is really a problem with adding this templates to CloudStack, wrong format or something like that. |
|
Hi, I tested on actual Host and i got the issue I have not seen this issue in HCI (uses 4.19 cloudstack) and the virter env! I don't know why but two time when i deployed it in physical host it had same issue. The other issue i reported of .img of ubuntu that comes next when system vm's will be resolved. It is not there in HCI but it is in virter. In short 
Logs as i added only one node and it was also the cloudstack management so. total 8.0K |
0 replies
|
linstor err l , Is there any error report available? We can check the content. |
0 replies
|
i have attached the logs already and other logs |
0 replies
|
Hi any update would be appreciated! |
3 replies
|
I thought we "solved" this and concluded that the wizard uses some how the wrong image template format? |
|
solved where, @rp- ? is there and issue or PR for it? (is this the qcow vs. raw issue?) |
|
Well I'm asking if the issue is just, that the wizard selected ubuntu templates are not working. If so this can be easily worked around by manually adding them correctly. |
|
Issue
I am more concerned toward the first issue as after fresh cloudstack + linstor deployment doesn't work on baremetal (system vm's issue). The same worked on virter (virtual env).
The ubuntu images normally comes in .img extension and not in qcow2 extension officially. |
1 reply
|
1 and a half problems here with the systemvm loop:
|
Ok i will try that and let u know. Thanks for responding. |
0 replies
|
Nope same thing even if i set fqdn everywhere 
tar cvpzf ~/cskvm01.logs.tar.gz /var/log/linstor-satellite/ /var/log/linstor-controller/ /var/log/cloudstack/ /var/log/libvirt/ /var/log/syslog /var/log/dmesg /var/log/kern.log |
1 reply
|
but the linstor node names are still You could use the But you would still need to fix the broken system systemvm template in Linstor, or recreate the CloudStack cluster. linstor spawn cloustack cs-1b41101f-f69f-11f0-a88f-de01a885d20e 5G
# qemu-img convert systemvm.qcow2 into /dev/drbd/by-res/cs-1b41101f-f69f-11f0-a88f-de01a885d20e as raw |
|
Nope i recreated everything now below is fresh install again.. 
also did below qcow2 to raw conversion here is the entire install script also here is some other things also i was using full fqdn and also made sure it the norm across but as u can notice linstor n l will show it first but later when adding storage it changes to short name. as you can see even though i added fqdn later it became shortname logs - I also tried removing and readding the primary storage you can check in log. also if i remove and add nfs primary it will start working but linstor primary doesn't that is the the issue still not resolved. |
1 reply
|
This makes no sense? See also a bunch of these errors: You can't have different node names on the same linstor controller. How did you try to rename the nodes? There is no rename functionality in Linstor. I would suggest you start from scratch and really only use the short hostname WITHOUT fqdn, as I think Cloudstack will remove the full qualified part for host searches. So use the cskvm01 notation everywhere. Also i should probably haven't suggested manually fixing the template resource, I think an easier approach would have been, to just delete and readd the Linstor primary storage, because that would have done the same. |
|
Also additionally here is the data dump of mariadb cloud db |
0 replies
|
Hi, This time i reinstalled the os and all. This time it worked. NO FQDN used anywhere. Systemvm's I think this was the issue the host file and the resolv.conf which i was overwritting. Thanks @rp- Also please look into the .img ubuntu tempaltes. that is still a issue. |
0 replies
|
update: |
0 replies
Answer selected by
DaanHoogland
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
update:
Hi the .img also works without issue so in total everything works