So to any sysadmins who do this kind of stuff on a daily basis, this is going to seem way obvious, but for somebody who doesn’t and has been struggling with this literally for months … let’s just say I’m pretty happy to finally have figured this out!
Also, this post is mostly for documentation’s sake so that I have a place to look back to when I need to do it again sometime many moons into the future…
It’s hard to believe that it’s been over a year already since I migrated my Plex server off of my old desktop hardware over to a proper rackmount server. Or at least Plex itself migrated, while the bevy of hard drives that 50+ TB of media lives on still resides in that aged and ever-waning PC.
Anyways, last June when I made the big leap to server-grade hardware, I only had a single hard drive to run VMs from for the new machine. For simplicity’s sake, I set it up as a RAID 0, single disk array, with the understanding that I could “easily” add more disks a few months later and re-configure that array into a more resilient RAID 5.
In fact, according to Amazon I did buy two more drives to use for said purpose in September 2018.
And just yesterday I finally got them working!
You see, it was probably too easy for me to setup that initial RAID 0 array via the new server’s BIOS. At the time, it seemed simple enough to add more drives to the pool and then reconfigure the array itself.
But one thing I’ve learned somewhat painfully since I first set this server up is that everything is more picky than that. Versions have to line up with the hardware, and older versions lack features supported by newer versions, even while they’re all being supported by the companies in parallel. This isn’t really news to me, but it’s certainly something that I never had to scrutinize to this extent.
With my old desktop server…
- Connect new hard drive.
- Find it in the CentOS Disks GUI, quick format it, and mount it.
With my new server…
- Connect new hard drive.
- Try to add it to my RAID pool via the RAID controller, but you can’t.
- Try to add it via ESXi, but you can’t.
- Try to connect via Dell OpenManage, but I didn’t install the server-side software in ESXi right because Dell’s support page for this server only goes up to ESXi 6.0 even though I’m running 6.5 and then I finally find the right software on a support doc found via Google.
- Try to connect via Dell OpenManage, but they only make a Windows client so I have to find a laptop to do that.
- Try to connect via Dell OpenManage, but the server doesn’t have a certificate and the login failure doesn’t mention that this is a big deal, so you just guess until you see a checkbox mentioning ignoring that and finally it works!
- Add new disks to RAID pool and reconfigure from RAID 0 to RAID 5 … and wait a very long time.
- Worry that I didn’t make backups of my VMs because I couldn’t figure out how to do it precisely the entire time.
- Try to expand virtual disk via ESXi now that the extra space is available, but it still doesn’t see it.
- Confirm via Dell OpenManage that the reconfigure is definitely done now and showing the extra space as available.
- Wait until 1:30am when nobody is using Plex and just reboot the whole thing, just in case.
- Try to expand virtual disk via ESXi, and now it sees it!
- Allocate additional space to new VM and reboot that VM, but it does nothing.
- Spend an hour Googling for instructions about how to allocate the new space inside of the guest OS until I finally found this random support post that ends up working not unlike magic!
- Verify that the new disk space is finally ready to use in the VM, and then debate whether it’s going to be enough or if I should’ve bought yet another disk just in case…
I mean, looking back logically it does make sense – first add the physical drives, then add them to the RAID pool, then rebuild the RAID array, then add new space to the Virtual Disk, then allocate the new space to a specific VM, then update the VM to recognize its new resources … maybe I was just hoping it would be slightly more seamless, even if only in parts! 😛
If anything, I guess it should be a tad easier the next time around, and now that I’ve gotten the bugs worked out of OpenManage, that alone is one less headache to worry about.
That said, I don’t want to rely on my work laptop for managing this server (and others down the road) indefinitely, so it also means I need to put together some sort of Windows box to sit in the corner and collect dust until it’s needed once in a blue moon…
Still, my Plex environment … minus the media itself … now lives on a cushy, new RAID 5 array that could sustain a single disk failure without missing a beat, plus I’ve got some extra cushion for downloading new stuff to boot.
Not too shabby for only ten months worth of work!