Help cutting my FreeNAS teeth

[gpsguy]

Don't worry about this now - as I said in an earlier post, you need to be concerned about the health of your drives. It appears that there's a problem with ada0. I'd get a replacement ASAP.

So I transferred TO my SSD using a WD Green via SATA at: 270MB/s max, ~120MB/s average. Total time: ~second for a single 8.73GB file.

Before you replace the drive, you'll need to follow the directions to do so. Many users don't read the instructions and create additional problems.

 

[cyberjock]

You have 4 disks.. can you post the smart data for the other 2 also please?

 

[victorzamora]

Cool. Well, I pulled the drive that I was thinking was bad. I have deleted the old volume and am about to create a 3-disk RAIDZ volume with the other 3 disks in its place.

 

[victorzamora]

About replacing the drive: my setup is new, and just pulling it out doesn't really hurt my feelings. I'm currently operating on a single 4TB partition, so a 3x2TB RAIDZ setup gives me what I had. Plus, I was at like 2.2TB....so no major worries. My HDD is out of warranty (I purchased an external hdd that was cracked open to get at the internal/bare HDD) and I don't feel like purchasing another. So for now, I'll stick with a 3x2TB RAIDZ array.

Since I already pulled that disk, I'm afraid the others have changed order (like ada1 isn't necessarily ada0 now) so I'm including files for all 3 of them.

 

[victorzamora]

Before you replace the drive, you'll need to follow the directions to do so. Many users don't read the instructions and create additional problems.

I'm sorry for having already pulled the drive out, I didn't read this before I did. Like I said, I'm not worried about losing data on the NAS yet as I haven't really put anything on it. I hope that this suggestion was simply to keep me from losing data.

Questions on that: I currently have another good 2TB WD Green....but it has all of my data on it.
1) I can't put that in my NASBox and "integrate" it dynamically into the other 3 disks for a 4-disk RAIDZ array, can I?
2) Could I reinsert that bad drive, create the volume, then pull the bad drive, transfer the data to the 3 good drives (4-disk RAIDZ volume with one disk pulled) and then "replace" the bad drive with the drive that currently contains my data?

Both of the above are long-shots, but I'm trying to think outside the box....and I'm thinking that just might fool the system. However, I'm risking a lot of data that I don't care to risk if I go either of those routes, correct?

 

[victorzamora]

Simply with the changes I've made I'm getting transfer speeds of ~93MB/s. That's a HUGE boost. The 60 minute transfer now took ~2 minutes. It's slightly slower than SATAHDD-to-SATASSD.....but good enough for me. Before I put my setup into the process of transferring my ~2TB of data over....should I run any more tests? Also, are my two ideas above capable of working or should I skip it and in turn skip the risk? I can always use the other 2TB drive for TimeMachine of two laptops and backing up my desktop, so I won't be too torn up about it :D

 

[cyberjock]

RAIDZ is "dead", just like RAID5 is "dead". RAID5/RAIDZ is not adequate protection for your data in today's world. Both were "dead" back in 2009...

http://www.zdnet.com/blog/storage/why-raid-5-stops-working-in-2009/162

 

[cyberjock]

Also, your Seagate ada2 disk has a non-zero value for current pending sector count and offline uncorrectable.

Maybe now you understand why RAIDZ2 is really the minimal protection you should use these days. It's also why I never recommend RAIDZ1..ever.

 

[victorzamora]

Wait, does that mean that the Seagate is going bad as well? If so, I can do a little bit of playing around to get 3 identical WD Green drives in there. The one I'd be replacing the Seagate drive with is a brand new WD Green 1x2TB drive.

 

[gpsguy]

BTW, did you change the idle setting on your WD Green drives? If you want to avoid a shortened drive life, then you ought to change the setting.

Read about it here: wdidle & WD Green drives

I've seen recommendations to set it to 300 seconds.

 

[victorzamora]

No I haven't. I'll be doing that once I get everything else handled. Thank you for the heads up.

 

[victorzamora]

Just a quick edit: My current plan is to use my Seagate as temporary storage. I have a new WD Green 2TB with all of my old data on it that I'll be transferring to my Seagate. I'll then use the new WD Green along with the other two (one old, one new, both verified "good" by the smartctl test) in a RAIDZ array. The Seagate will become the backup location for a couple of my laptops. My laptops have no critical data on them, it's mostly just settings I'd hate to lose. I'll be replacing it soon(ish).

 

[gpsguy]

I'd also suggest you run "long" tests on each of your drives, before you start to use them in your NAS. It'll probably take a couple of hours. As always, substitute a device number for the "x".

smartctl -t long /dev/adax

 

[victorzamora]

Okay, so....what will the output look like? Should I save it to a txt file? What will tip me off to the fact that it's good or bad or going bad? If it's going to take several hours per drive, I want to know it's worth it....y'know?

 

[gpsguy]

Here are some examples: test-a-hard-disk-with-smartctl. Yes, you'd want to capture the output to a file.

Running S.M.A.R.T tests should be part of your routine maintenance. See section 4.5 in the manual.

Is it worth doing this work? Do you check your oil in your car? How about the tires?

Had you done these simple tests first, you would known about the drive issues from the start.

 

[victorzamora]

Thanks for the link. I'll start making that part of my routine. How often do you recommend I do it? I'll read that link, read up what I can on further SMART testing, and try to polish of the manual sooner rather than later. Looking back on it now, having run those tests would have helped save a tremendous amount of time.

One question I haven't actually found the answer to is: Why use Putty? As far as I can tell, it's a way to control the NASBox in DOS without going through the Web UI. I'm sure it's better/safer/easier.....but I can't imagine that's the only reason to use it.

 

[gpsguy]

The FreeNAS documentation suggests scheduling a long SMART test "once or twice a month".

Why use PuTTY? I consider it a much better way to interact with our FreeNAS servers, 99% of the time. With it, one can easily redirect session output to a file. I thought about giving you the information that cyberjock furnished about redirecting it to a file with a ">", but felt that might lead to as many questions, as would using PuTTY.

I can set the session logging to "printable output" and capture all this command line information to a file. I can edit it later, incorporate it into my documentation, and could post it here, if need be. One can also paste text into the console via PuTTY. There might be times, when it's advantageous to be able to do so.

Many of us run our servers headless, ie. no monitor or keyboard. I don't remember when they added the ability to access the console via the Web UI, but it was a new feature in the year or so. Prior to that, one could only access it via SSH or the console. In 1% of the time, one might need to hook up a monitor and keyboard, if the server won't boot and you can't get to it otherwise. In that circumstance, a camera pix, might be the best information you can provide.