Huh, interesting. The wearout of my two Proxmox-SSDs are 76% and 81%...I think it might actually be the time to replace them :D Thanks for the video! I now know more then I did before watching the video!
Might make sense to install rasdaemon for logging MCEs and ECC errors (note that the version you get from apt does not work correctly. It has to be an install from source)
Thanks for this video. It was very informative. What hardware are you running on that box? I noticed 1235U CPU but also a bunch of spinning drives, so seems like a neat relatively low power node.
It's a very interesting video. How about one on disaster recovery such as an upgrade failure and need to install the latest version to an existing series of pools or loss of access to pools by users or making the individual nodes fault tolerant?
Glad you liked the video. Disaster recovery is hard to make a video about as there are many weird ways things can go wrong, but I’ll try to cover this more in future videos.
Can You make a video Using Turnkey Fileserver in ProxMox..specifically adding a Hdd with media already on it…also do you have to do a pass thru in proxmox before using Turnkey FileServer
Yea it’s been a while since I looked at turnkey so I should look at it again. As far as using disks with existing data you likely want to mount the drive in proxmox host, then map the mount point of the drive to a container.
Here is my quick guide. I'd also typically reccomend moving the data to a linux native filesystem, but depends on the exact setup. This guide is for mounting the drive in the host and passing the mount point to a vm. You can also pass a whole drive to a vm. -Mount the drive on the host. Look into guide for mounting drive via fstab for how to have it auto mount at bootup. -Pass the drive to the container. Either edit the config file or run pct set $VMID -mp0 /drive/mount/on/host,mp=/path/to/mount/in/ct. Then boot the container and it should be accessible at the /path/to/mount/in/ct you set.
Hi Brandon, I have a wearout of 34% on my primary Proxmox-Server SSD, should I be worried? (SMART Test is ok!) Could you point me to (perhaps) one of your vids for replacing (without much hassle) the boot SSD. I have two Mini-PCs for my Proxmox servers, both AMD and not with Enterprise SSDs. I could either migrate all VMs and CTs to my secondary server or restore from PBS backup. I really appreciate your work and would love to get any hints! 🙏
I really need to make a video about replacing mirrored ZFS boot drives. If your boot drive is of a different config I'd try cloning it over via DD or other tools via a live disk, but a backup restore will also work fine. With 34% wearout I wouldn't be worried, but maybe take a look at the rating in a month or two to see how its changing to get a estimate of time remaining.
@@ElectronicsWizardry Thank you so much for your quick answer. I‘m not really good in CLI commands, I know that there is this Clonezilla tool, but never experienced with it. Backup/Restore would be easy to do for me, but I would prefer to clone, although I would then be required to resize somehow the partitions, as the new drive would be a 2 TB vs a 500GB now. Looking forward to a video as you mentioned. Greetings from Luxembourg 🇱🇺.
@Mr.Jean-Paul is be tempted to give clonezilla a shot and see how it goes. By default the extra space of the 2tb drive should be unused. Then I’d use a Linux live disk with gparted to expand the partitions. This can be complex depending on how the storage is setup. If you haven’t done this much before I’d probably go with the safe but more time consuming approach of making and restoring backups.
I think those NVMe drives I had laying around that were included with miniPCs. Not great drives, but they work. Its annoying that the model reported for SSDs is often different than the advertising name.
Is it possible to fine tune what we get for email notifications in proxmox, like adding email notification for a specific error or turning it off for that specific error? for example disk wear reached x%. Is there a built-in or third party tool for this? A management interface would be nice for this.
Proxmox shouldn't kill a VM that is running, but there are enough things that can happen that would cause a VM to shutdown that I'd check it to make sure nothing weird is going on. OOM errors on Proxmox can kill vms without errors or other crashes, or a guest shutting its self down(I've seen licensing errors cause this in windows)
@@ElectronicsWizardry Oooh...okay. Gotcha. But you'd think that is, in a way, an issue with the hardware that the Proxmox server is running -- in the sense that if you're getting OOM issues, then it may mean that the host itself needs more RAM, or needs to find a way to slim down the VM's memory footprint/requirements.
@@ewenchan1239 Yea VMs that are shutdown other than by a user could be from various reasons that would be worth looking into. I'd like it Proxmox could better log if VMs are forced killed by the host or crash to better track down issues when they happen.
@@ElectronicsWizardry This is one of the reasons why for Linux clients/VMs, I've switched them over from VMs to LXC containers instead. Unless there is a very specific reason why I would actually need a Linux VM, otherwise, I have found that in terms of resource usage, LXC containers are more efficient from that perspective. (Especially since you can just specify the MAXIMUM amount of RAM that the LXC container is allowed to use, but doesn't allocate that amount of RAM initially, like how a VM would "carve out" RAM from the Proxmox host, for the VM.) And the only time that I need a Linux VM is if I want to run multiple monitors on said Linux VM and/or if I want to use virtiofs on Proxmox for VM host data communications.
Huh, interesting. The wearout of my two Proxmox-SSDs are 76% and 81%...I think it might actually be the time to replace them :D Thanks for the video! I now know more then I did before watching the video!
Might make sense to install rasdaemon for logging MCEs and ECC errors (note that the version you get from apt does not work correctly. It has to be an install from source)
Says the man who let 2 of his disks die in his Z2 ZFS Pool 😂 Keep up the good work.
I think the saying that a mechanics car doesn't run well also applies to IT people.
@@ElectronicsWizardry As we IT people say Its futureme's problem :)
Thanks for this video. It was very informative.
What hardware are you running on that box? I noticed 1235U CPU but also a bunch of spinning drives, so seems like a neat relatively low power node.
I was reusing the Ugreen NAS for this video. The i5 1235u is a mobile i5(2p + 8e) and has 6 3.5in hdd bays.
It's a very interesting video. How about one on disaster recovery such as an upgrade failure and need to install the latest version to an existing series of pools or loss of access to pools by users or making the individual nodes fault tolerant?
Glad you liked the video. Disaster recovery is hard to make a video about as there are many weird ways things can go wrong, but I’ll try to cover this more in future videos.
Can You make a video Using Turnkey Fileserver in ProxMox..specifically adding a Hdd with media already on it…also do you have to do a pass thru in proxmox before using Turnkey FileServer
Yea it’s been a while since I looked at turnkey so I should look at it again. As far as using disks with existing data you likely want to mount the drive in proxmox host, then map the mount point of the drive to a container.
@@ElectronicsWizardry Thank You for responding do you have a step by step?
Here is my quick guide. I'd also typically reccomend moving the data to a linux native filesystem, but depends on the exact setup. This guide is for mounting the drive in the host and passing the mount point to a vm. You can also pass a whole drive to a vm.
-Mount the drive on the host. Look into guide for mounting drive via fstab for how to have it auto mount at bootup.
-Pass the drive to the container. Either edit the config file or run pct set $VMID -mp0 /drive/mount/on/host,mp=/path/to/mount/in/ct.
Then boot the container and it should be accessible at the /path/to/mount/in/ct you set.
@@ElectronicsWizardry A video about this would be great! Keep up the good work, the content is tits
Hi Brandon, I have a wearout of 34% on my primary Proxmox-Server SSD, should I be worried? (SMART Test is ok!)
Could you point me to (perhaps) one of your vids for replacing (without much hassle) the boot SSD. I have two Mini-PCs for my Proxmox servers, both AMD and not with Enterprise SSDs. I could either migrate all VMs and CTs to my secondary server or restore from PBS backup.
I really appreciate your work and would love to get any hints! 🙏
I really need to make a video about replacing mirrored ZFS boot drives. If your boot drive is of a different config I'd try cloning it over via DD or other tools via a live disk, but a backup restore will also work fine.
With 34% wearout I wouldn't be worried, but maybe take a look at the rating in a month or two to see how its changing to get a estimate of time remaining.
@@ElectronicsWizardry Thank you so much for your quick answer. I‘m not really good in CLI commands, I know that there is this Clonezilla tool, but never experienced with it. Backup/Restore would be easy to do for me, but I would prefer to clone, although I would then be required to resize somehow the partitions, as the new drive would be a 2 TB vs a 500GB now. Looking forward to a video as you mentioned. Greetings from Luxembourg 🇱🇺.
@Mr.Jean-Paul is be tempted to give clonezilla a shot and see how it goes. By default the extra space of the 2tb drive should be unused. Then I’d use a Linux live disk with gparted to expand the partitions. This can be complex depending on how the storage is setup. If you haven’t done this much before I’d probably go with the safe but more time consuming approach of making and restoring backups.
@@ElectronicsWizardry Agreed! Backup and restore…will choose the safe way. Greetings
What models are those 2 nvme disks?
CPO41BY1599326
YSO128GTLCW-E3C-2
I think those NVMe drives I had laying around that were included with miniPCs. Not great drives, but they work.
Its annoying that the model reported for SSDs is often different than the advertising name.
Is it possible to fine tune what we get for email notifications in proxmox, like adding email notification for a specific error or turning it off for that specific error? for example disk wear reached x%. Is there a built-in or third party tool for this? A management interface would be nice for this.
I don't know of a easy setup off the time of my head for this, but I need to do a video looking into email notifications soon.
wait...why would Proxmox randomly kill VMs that are supposed to be running???
Proxmox shouldn't kill a VM that is running, but there are enough things that can happen that would cause a VM to shutdown that I'd check it to make sure nothing weird is going on. OOM errors on Proxmox can kill vms without errors or other crashes, or a guest shutting its self down(I've seen licensing errors cause this in windows)
@@ElectronicsWizardry
Oooh...okay.
Gotcha.
But you'd think that is, in a way, an issue with the hardware that the Proxmox server is running -- in the sense that if you're getting OOM issues, then it may mean that the host itself needs more RAM, or needs to find a way to slim down the VM's memory footprint/requirements.
@@ewenchan1239 Yea VMs that are shutdown other than by a user could be from various reasons that would be worth looking into. I'd like it Proxmox could better log if VMs are forced killed by the host or crash to better track down issues when they happen.
@@ElectronicsWizardry
This is one of the reasons why for Linux clients/VMs, I've switched them over from VMs to LXC containers instead.
Unless there is a very specific reason why I would actually need a Linux VM, otherwise, I have found that in terms of resource usage, LXC containers are more efficient from that perspective.
(Especially since you can just specify the MAXIMUM amount of RAM that the LXC container is allowed to use, but doesn't allocate that amount of RAM initially, like how a VM would "carve out" RAM from the Proxmox host, for the VM.)
And the only time that I need a Linux VM is if I want to run multiple monitors on said Linux VM and/or if I want to use virtiofs on Proxmox for VM host data communications.
you have good eyes and jaw line and eye brows , why not shaving that hair off ?