Unraid Hardware - What did you do?

[Taddeusz]

Originally Posted by Fuzzy

I can say one thing... get a case that has easily accessible drives.. doesn't have to be hot-swap bays or anything fancy, but even some of the nicer towers with the full-height drive cage with trays would be great. Just noticed a drive had failed (7 years old, so I don't feel that bad) while I was on vacation. Got home, ordered a new drive from amazon. It arrived this morning, and I was able to stop the array, power down, replace it, and power back up in about 15 minutes (label the trays with which unRAID disk# they are holding). Kick off the rebuild (which will take like 9 hours I think), and get right back to watching tv.

Total downtime for a drive failure with zero data loss 15 minutes... gotta love unRAID!

Yeah, I've had to replace a couple drives since I started running unRAID. It's a wonderful thing. Especially with the Docker & VM functionality.

[Fuzzy]

Originally Posted by Taddeusz

Yeah, I've had to replace a couple drives since I started running unRAID. It's a wonderful thing. Especially with the Docker & VM functionality.

Probably the greatest thing was being able to choose the downtime. The drive failed a week ago, and didn't affect operation of anything until i took the time to shut it down and replace it this morning. I'm thinking about throwing my new 'ready spare' in as second parity volume, but I don't really know if there is the easy ability to easily transition a second parity to a data drive.

[BobPhoenix]

Originally Posted by btrcp2000

How did you know this about pci? Need to make sure the E5 2660 v2's (sr1ab) I am considering will handle passthrough

Looks like your sr1ab should be ok based on page here (VT-d is mentioned):
http://www.cpu-world.com/sspec/SR/SR1AB.html

When you look at the SR0H8 VT-x is mentioned but not VT-d:
http://www.cpu-world.com/sspec/SR/SR0H8.html

[glenner]

Originally Posted by Fuzzy

I can say one thing... get a case that has easily accessible drives.. doesn't have to be hot-swap bays or anything fancy, but even some of the nicer towers with the full-height drive cage with trays would be great. Just noticed a drive had failed (7 years old, so I don't feel that bad) while I was on vacation. Got home, ordered a new drive from amazon. It arrived this morning, and I was able to stop the array, power down, replace it, and power back up in about 15 minutes (label the trays with which unRAID disk# they are holding). Kick off the rebuild (which will take like 9 hours I think), and get right back to watching tv.

Total downtime for a drive failure with zero data loss 15 minutes... gotta love unRAID!

Fuzzy... that's a happy story!

I'm just curious what does a drive failure look like in this case when you come back from vacation? The drive is "red balled" in the dashboard and unraid sends you some kind of notification?

Does the SageTV docker continue recording shows (hugely IO intensive) and unraid continue processing while the drive has failed and until you replace it? I'm not quite clear yet how I will handle a drive failure when it finally occurs, but good to hear the solution can be painless. I would tend to think I would want to shut down as many dockers and VMs as possible until the drive is replaced and restored with the parity rebuild.

Numbering the drive bays to match the unraid setup is a good idea. I had already done that. I also have some screencaps of my unraid dash board to just document my basic config.

I got the Fractal Design R5 mid tower case. Fits 8+ internal drive bays and they are very easy to get at or replace. Well designed box.

[Fuzzy]

Originally Posted by glenner

Fuzzy... that's a happy story!

I'm just curious what does a drive failure look like in this case when you come back from vacation? The drive is "red balled" in the dashboard and unraid sends you some kind of notification?

Does the SageTV docker continue recording shows (hugely IO intensive) and unraid continue processing while the drive has failed and until you replace it? I'm not quite clear yet how I will handle a drive failure when it finally occurs, but good to hear the solution can be painless. I would tend to think I would want to shut down as many dockers and VMs as possible until the drive is replaced and restored with the parity rebuild.

Numbering the drive bays to match the unraid setup is a good idea. I had already done that. I also have some screencaps of my unraid dash board to just document my basic config.

I got the Fractal Design R5 mid tower case. Fits 8+ internal drive bays and they are very easy to get at or replace. Well designed box.

Nope, only indication was the drive listed as missing. I never got around to setting up any sort of notification, so that's really all I got. Sage had no idea anything was amiss, and recordings just kept on keeping on. Even streamed quite a bit while on vacation, and things were just fine.

[graywolf]

There is a unRaid plugin (Server Layout) that I like. Keeps the info on what drives are where in my 24 drive 4u server. Have to get around to it (pull the old paperwork) to put in the warrenty info, date put in service, etc.

[MacDaddy]

If your case has drive cages that have an exposed activity LED for the drives, then this link has a utility script that will help identify the physical drive associated with the proper mount. https://forums.lime-technology.com/t...comment=167215

[glenner]

Originally Posted by Fuzzy

Nope, only indication was the drive listed as missing. I never got around to setting up any sort of notification, so that's really all I got. Sage had no idea anything was amiss, and recordings just kept on keeping on. Even streamed quite a bit while on vacation, and things were just fine.

Interesting... I wonder if that "bad" drive is mountable... if so, you may be able to look for any SMART errors. Specifically the "reallocated sector count" is the one that is supposed to indicate a drive failure is imminent.

The notification framework on unraid actually looks pretty useful to me, and if it works is clearly a step up from my old Windows setup. As best as I can tell, I'm expecting to get email messages if any of my drives start to show SMART errors. That may allow me to proactively replace a drive before it fails.

We'll see...

[Fuzzy]

Originally Posted by glenner

Interesting... I wonder if that "bad" drive is mountable... if so, you may be able to look for any SMART errors. Specifically the "reallocated sector count" is the one that is supposed to indicate a drive failure is imminent.

The notification framework on unraid actually looks pretty useful to me, and if it works is clearly a step up from my old Windows setup. As best as I can tell, I'm expecting to get email messages if any of my drives start to show SMART errors. That may allow me to proactively replace a drive before it fails.

We'll see...

No, the bad drive was well and good gone.. no response to the SATA port on the thing, without even BIOS seeing it (electronics failure, not platter or head issues). No matter, it was old, and i had parity. Was a good reason to replace it with a bigger and faster drive and carry on.