Proxmox containers won't start but VM hosts start fine.

[dsexton18]

I am a bit perplexed. I have proxmox 8.2.4 installed. After I create a new container and try to start it the container never starts it's stuck in an endless start mode.
VMs start fine both VM and containers are are on the same storage.

I have my proxmox host connected direct to my synology nas over 10GB connection.

I can see when I start a container the IO delay is around 6%.

If I reboot the proxmox host vms start fine and no IO delay. What gives?

I am using Minisforum MS-01 for my proxmox host.

 

[sw-omit]

If you migrate the container to / make a new test-container on local (thin-)lvm storage to test, does it then boot normally?

 

[dsexton18]

If you migrate the container to / make a new test-container on local (thin-)lvm storage to test, does it then boot normally?

Containers created on local-lvm boot fine no issues or migrated to local-lvm work fine.

 

[sw-omit]

Ok, so we at least know it is something related to the (remote) storage and containers.

I'm not that well versed with containers, so I probably can't help you much further, but at least one more thought:

How is that NAS connected software-wise?
SMB / NFS / iSCSi (/other) ?
I know that Synology has all those options on its nas, have you tried setting up a (temporary) share with one of the other protocols and testing it on that?
Would help to both further drill down to where the issue is, and perhaps that other connection-method is "good enough" (for your Containers)

 

[dsexton18]

It's a samba share. Funny things is I initially had my containers on a smb share no issues but via 10gbe. I created a new share that's I am using currently. I just updated the ip in /etc/pvc/storage.conf. initially the smb share was added with my ip that was a 1gbe connection by mistake. I updated the ip to use the 10gbe connection.

 

[sw-omit]

Since NFS and SMB/CIFS have the same capabilities [1] and is also fairly easy to set up on Synology's side [2], , but do use a different plugin to do the talking, it could be an option to test out at least I would say. Do note, while you could re-use the same share, I would suggest to use a separate share for testing.

That said though, for a more in-depth testing of why it doesn't work on SMB for containers (in your case), you might have to wait for someone else to chime in with things to check.

[1] https://pve.proxmox.com/wiki/Storage#_storage_types
[2] https://kb.synology.com/en-br/DSM/t..._on_Synology_NAS_within_the_local_network_NFS

 

[dsexton18]

I am thinking my issue may of been my SFP cable I ordered 10Gtek SFP+ DAC Twinax Cable off ebay. Either the cable was bad or perhaps I damaged it. Now I can start containers but I have noticed after the container is created I have to wait 3 or 4 seconds before I start it. Otherwise I get

()
TASK ERROR: volume 'pve3_lxc:9100/vm-9100-disk-0.raw' does not exist

more testing to come.

 

[dsexton18]

I added an NFS share and can start/stop containers so far no issues using NFS storage. Seams like my issue is related to smb/CIFS.