I’m trying to set up a PMM server using the virtual appliance.
The VM is set up and I can ssh on, but there are no pmm processes visible and no postgres database set up, for example.
I’ve set up a postgres db and tried running pmm-managed manually which is failing with an error regarding database migration due to authentication issues despite having verified the user and access (using md5 not peer authentication and passing in the password).
We’re using KVM so had to convert the OVA file, but I’m not aware of any problems in the creation of the VM itself.
Reading the documentation it suggests the virtual appliance should ‘just work’ once the VM is spun up, but it doesn’t seem to be the case here.
Am I missing something?
Is docker running inside the VM? I’m pretty sure that’s all the OVA does: launch docker and run PMM inside docker. If not, honestly, I would just abandon the OVA and setup a regular linux VM and run the 2 docker commands to install PMM.
Thanks for that. I had assumed the OVA built a ‘dedicated server’ running PMM, not just spin up a docker container.
I’ve had some intermitted network issues with my containerised instance so was hoping the virtual appliance would be more reliable.
I’ll rebuild the docker instance and go with that. Thanks
What version of PMM are you running? What we currently release as OVF is an OS with PMM running directly on it (NOT nested inside docker). There are newer experimental versions we’re testing out that might adopt that model (OVF as a bare bones OS with Docker running PMM but we’re still doing performance comparisons to make sure there’s no impact to the end user).
I’ve never converted the OFV for KVM but I’m running 2.27.0 on Vmware ESXi with no issues. You should be able to run
supervisorctl status to see if any of the required processes are running. You can also look in
/srv/logs/ to get a feel for what might be failing to start and why. Here’s a screenshot of what’s running on my instance to compare on your own…so if grafana or postgres failed to start you can look at the particular log to see why.
We’ve already pulled the Vm down and just thrown a regular VM back up to run Docker inside it.
I’ll maybe try doing it locally and see if I can get it working that way and then go backl to the original box.