Some Basic Questions - Upgrading Drives

Hi unRAID community. I am in the process of upgrading several drives in my server. I have 3 5TB Reds that are ready to go in, while my server has several 1TB and 2TB drives. I have read the following instructions for upgrading the Parity drive (which I will do first) as well as the process in replacing data disks:

 

https://lime-technology.com/wiki/index.php/The_parity_swap_procedure

 

https://lime-technology.com/wiki/index.php/Replacing_a_Data_Drive

 

I have several questions:

 

1. Since I want to replace these drives physically, I want to be able to make sure I remove the old drives from their original SATA ports. However, I want to re-add them in a different port. If I reconnect them, is there anything extra that I need to do before or after I connect them> I'm only confused because unRAID undoubtedly will recognize them as a drive that was previously in my system so I'm not sure what will happen if I reconnect them.

 

2. Something to do with my LSI 9211-8i card, I guess the right place to look is here: http://lime-technology.com/forum/index.php?topic=12767.0 ** Place holder for my question **

 

Thank you.

Forget about the parity swap procedure. That is only needed when you have a failed data drive and you want to rebuild it to the current parity drive while upgrading parity to a larger drive. Basically it copies parity from the old parity to the new parity drive and then rebuilds the failed data disk to the old parity. It is unnecessary complication in your situation. Just start with the parity upgrade by replacing it first and rebuild parity.

 

Are you wanting to rebuild your existing drives to larger drives or just add larger drives to your array? I am a little unclear what you have in mind with this part.

... Since I want to replace these drives physically, I want to be able to make sure I remove the old drives from their original SATA ports. However, I want to re-add them in a different port...

Do you want the old drives to be part of the array with their data intact? Or do you want to rebuild their contents onto larger drives and add them back later without any data? The specific ports aren't important to unRAID since it keeps track of which drive belongs in which array slot by each drive's serial number.

 

In any case, I assume you know that you can only replace/rebuild one disk at a time.

Thanks for that. I've gone ahead with the upgrade of the parity drive and it had 2 hours left (15hrs total) in rebuilding. I intend to replace the data drives with the larger ones, then reassign the older and smaller drives with different shares. So I still want them to be part of the array but without the data.

 

Cool, thanks again regarding the ports. Yeah I am only installing one disk at a time at this stage.

 

Edit: I was getting invalid data warnings/errors initially after initiating a parity rebuild. I'm assuming these might be safe to ignore, I'll confirm when I get home (rebuild should complete by then).

I think you were probably just getting notified that you were in a non-protected state while you rebuild parity.

 

Maybe you already know, but if you add the old smaller drives to new slots in the parity-protected array you will need to preclear them first or unRAID will clear them with your array offline so parity will remain valid.

I think you were probably just getting notified that you were in a non-protected state while you rebuild parity.

 

Maybe you already know, but if you add the old smaller drives to new slots in the parity-protected array you will need to preclear them first or unRAID will clear them with your array offline so parity will remain valid.

 

Yeap, all good now. Got home and checked and got the status "Notice: Parity disk returned to normal operation"

 

Doing the rest of the 2x data drive replacements but I will only install one new drive at the time, but I'll keep the old ones (moving them down the SATA ports) and will pre-clear them after the new drive assignment is done.

 

Thank you.

 

Update:

 

Hooked up the new data drive (disconnected old one but still left it in chasis, waiting for my Mini SAS 36pin (SFF-8087) x2 cables so I can reconnect the old ones as they arrive), then assigned the new drive, and started the array, now I can't access it via the browser. I can still ping the server, Plex is even running, and I can even browse all the shares via Explorer (connecting over my network from a Windows 10 PC). I am assuming that it's just rebuilding that disk now, I should wait overnight?

Update:

 

Hooked up the new data drive (disconnected old one but still left it in chasis, waiting for my Mini SAS 36pin (SFF-8087) x2 cables so I can reconnect the old ones as they arrive), then assigned the new drive, and started the array, now I can't access it via the browser. I can still ping the server, Plex is even running, and I can even browse all the shares via Explorer (connecting over my network from a Windows 10 PC). I am assuming that it's just rebuilding that disk now, I should wait overnight?

A rebuild should not normally take the GUI offline. 

 

It might be worth getting the syslog by running the 'diagnostics' command from a console/telnet session and posting the ZIP file here (assuming the command completes OK) to see if anyone can spot anything.

Update:

 

Hooked up the new data drive (disconnected old one but still left it in chasis, waiting for my Mini SAS 36pin (SFF-8087) x2 cables so I can reconnect the old ones as they arrive), then assigned the new drive, and started the array, now I can't access it via the browser. I can still ping the server, Plex is even running, and I can even browse all the shares via Explorer (connecting over my network from a Windows 10 PC). I am assuming that it's just rebuilding that disk now, I should wait overnight?

A rebuild should not normally take the GUI offline. 

 

It might be worth getting the syslog by running the 'diagnostics' command from a console/telnet session and posting the ZIP file here (assuming the command completes OK) to see if anyone can spot anything.

 

I feel a bit silly because it doesn't look like I have the diagnostic tools, I don't remember downloading it from the Tools page. I ran it and it gave me a  "-bash: cd: boot: No such file or directory" and even when I check the logs directory there are no diagnostic zipped files in there.

 

I ran cat /var/log/syslog and here's the txt file. I turned the network port on the switch (web console) off then on again a few times, I get the problem sometimes which I've simply needed to do this to re-connect/gain access to the GUI.

Syslog.zip

The reason you don't have Diagnostics is you are on V5 and you have posted in V6 support. The basic way things work for your purposes haven't really changed though. Just had us confused for a while.

 

What do you get from the command line with this?

ps aux | grep emhttp

 

Moving this thread to V5 area.

The reason you don't have Diagnostics is you are on V5 and you have posted in V6 support. The basic way things work for your purposes haven't really changed though. Just had us confused for a while.

 

What do you get from the command line with this?

ps aux | grep emhttp

 

Moving this thread to V5 area.

 

My apologies, I didn't check my version and assumed I was on the latest version.

 

As for the output of that command, I get the following:

 

root      3976  0.0  0.0  5764  1500 ?        Sl  Jan29  0:06 /usr/local/sbin/emhttp

root      7282  0.0  0.0  2452  584 pts/0    S+  00:25  0:00 grep emhttp

The reason you don't have Diagnostics is you are on V5 and you have posted in V6 support. The basic way things work for your purposes haven't really changed though. Just had us confused for a while.

 

What do you get from the command line with this?

ps aux | grep emhttp

 

Moving this thread to V5 area.

 

My apologies, I didn't check my version and assumed I was on the latest version.

 

As for the output of that command, I get the following:

 

root      3976  0.0  0.0  5764  1500 ?        Sl  Jan29  0:06 /usr/local/sbin/emhttp

root      7282  0.0  0.0  2452  584 pts/0    S+  00:25  0:00 grep emhttp

That seems to indicate that the webUI is still running.

 

You are probably 2 years from the latest version. A lot has happened. Even if you don't care about some of the new "running more stuff on your server" features V6 has a lot of improvements in the NAS functionality. And V6 is easier for us volunteers to support since we tend to stay on the bleeding edge. You should consider upgrading when you get everything squared.

Will do. I think the data rebuild should be done soon, I'm hoping I can get back in. Otherwise I'll check for disk activity via telnet and issue a shut down from there after checking to make sure there are no activities.

Update: I stopped the array and unmounted each drive. Disk4 was busy, used the guide to find what process and what user was doing this and it was actually the plex services. I stopped them and managed to unmount all drives followed by issuing a reboot. Post reboot I can get back to the interface again and all seems well. Before all this I checked CPU usage and it looked pretty idle, which is what I used to assume that the server was no longer doing the rebuild. My files look intact, I can see and access them no issues. I'll continue with one more disk upgrade, make sure that rebuild completes, run a parity check, then upgrade the server.

Update: I stopped the array and unmounted each drive. Disk4 was busy, used the guide to find what process and what user was doing this and it was actually the plex services. I stopped them and managed to unmount all drives followed by issuing a reboot. Post reboot I can get back to the interface again and all seems well. Before all this I checked CPU usage and it looked pretty idle, which is what I used to assume that the server was no longer doing the rebuild. My files look intact, I can see and access them no issues. I'll continue with one more disk upgrade, make sure that rebuild completes, run a parity check, then upgrade the server.

I would recommend doing a non-correcting parity check before proceeding with the next disk. The rebuild process writes but doesn't verify, it relies on the drive kicking out an error to determine if the write failed. Given the circumstances here, I wouldn't feel comfortable without verifying that the rebuild was in fact complete. The next disk rebuild will be corrupt if there was an issue with this one.

 

Also, I'd strongly recommend not writing to the server, and disabling any plugins while doing the disk rebuild.

Update: I stopped the array and unmounted each drive. Disk4 was busy, used the guide to find what process and what user was doing this and it was actually the plex services. I stopped them and managed to unmount all drives followed by issuing a reboot. Post reboot I can get back to the interface again and all seems well. Before all this I checked CPU usage and it looked pretty idle, which is what I used to assume that the server was no longer doing the rebuild. My files look intact, I can see and access them no issues. I'll continue with one more disk upgrade, make sure that rebuild completes, run a parity check, then upgrade the server.

I would recommend doing a non-correcting parity check before proceeding with the next disk. The rebuild process writes but doesn't verify, it relies on the drive kicking out an error to determine if the write failed. Given the circumstances here, I wouldn't feel comfortable without verifying that the rebuild was in fact complete. The next disk rebuild will be corrupt if there was an issue with this one.

 

Also, I'd strongly recommend not writing to the server, and disabling any plugins while doing the disk rebuild.

 

Done, thanks for that. Stopped Plex (only thing I have running) and just kicked off a non-write parity as you suggested. When this completes successfully I'll proceed to the next disk. Once the last disk is complete, and I'll do another non-write parity check, should I re-instate the older disks first or should I upgrade to version 6 THEN reinstate the older disks?

 

Update: Completed with 0 errors. 

Kicked off the final data drive replacement, exact same thing and killed the GUI. Ok now what to do now anyway. I'll update to version 6 as soon as the parity check is complete. I read somewhere there's a tool that I can run and do the upgrade.

 

Update: Wow, came home, 2hrs left, and surprisingly the GUI is back. 

 

Update 2: Great. Upgraded using the plg file and now I am getting a "Could not find kernel image: linux"

 

Update 3: Got it working, I had to copy over the syslinux from the 6 zip and run the make_bootable. Now painfully trying to restore the most important thing: My Plex server and settings... I feel like I'm not going to make it to work tomorrow morning.

You should definitely go with a docker for plex. You can probably get it to use the same plex library you had before if you set it up right then you won't have to rescan.

You should definitely go with a docker for plex. You can probably get it to use the same plex library you had before if you set it up right then you won't have to rescan.

 

Thanks for that, I read the guide and I'll do my best to configure it using docker.

 

I'm having weird issues with unRAID shares moving

https://lime-technology.com/forum/index.php?topic=46170.0

 

But I think I'll try and do a new config at this stage. It's not really necessary because all my data is there, it's just I want them organised a certain way. Thank you.