SOLVED Beginning Burn-in, help with SMART results?

[Fraoch]

Why not use tmux instead of trying to hack your way around?

I'm not sure if this was directed at me or at @J-NAS but I only saw about tmux after I'd finished badblocks tests, plus it looked a little tricky. I might want to learn tmux because I accidentally disconnected a session or two at the start of testing, necessitating starting all over again.

 

[Ericloewe]

I'm not sure if this was directed at me or at @J-NAS but I only saw about tmux after I'd finished badblocks tests, plus it looked a little tricky. I might want to learn tmux because I accidentally disconnected a session or two at the start of testing, necessitating starting all over again.

It does require some reading, but it's definitely the better option, given the alternatives.

 

[J-NAS]

*sinking feeling in gut*

This is going off the rails into areas I have not fully researched. I had glossed over tmux but had initially dismissed it as a bit much to chew if sticking to direct dd and badblock commands would have worked out in my favour.

Not asking to be spoon-fed, but is there a FreeNAS-centric tmux "guide" anyone could recommend? A good resource to wrap my head around how I would use it to perform the badblocks test(s)?
I'm assuming this guide will be sufficient...

I enabled SSH and logged in and used "top", but it seemed to kill my dd process. I saw it at the top of the stack, and then it evaporated. Confirmed its absence via the WebGUI version of show processes.

It's only been running for 1.5hrs, so I doubt it just happened to terminate co-incidently as I pulled it up with "top".

I was really hoping to get through the testing phase with the research I'd already performed. Doh! At least now I know how to enable SSH and login with it!

 

[J-NAS]

Yah, this is why I should do all my reading first. I read those instructions, executed tmux, and began running four panels of dd write tests. When I tried to created the fifth it told me it couldn't create the panel because it didn't have the space.

Reading the "man" page, it seems the quotation key is splitting the screen in half. Perhaps not the best way to get to six panels. I used "c" to create a new window to work on the remaining two drives. Seemed to work. Perhaps this will help some other soul in the future. Read the man page before you get too excited.

 

[J-NAS]

You can't reopen a closed SSH session and pick up where you left off, unfortunately, so tmux is the way to go. It also allows for parallel testing of all disks.

This confuses me. I am currently running tmux over SSH via putty. Am I safe to close down putty, login back in via SSH 12hr from now and execute tmux attach? Will that show the outcome of all 6 panels I opened? Do I have to do anything special beyond closing putty? Issue a detach command, key combination or something like that?

Thanks.

 

[Ericloewe]

This confuses me. I am currently running tmux over SSH via putty. Am I safe to close down putty, login back in via SSH 12hr from now and execute tmux attach? Will that show the outcome of all 6 panels I opened? Do I have to do anything special beyond closing putty? Issue a detach command, key combination or something like that?

Thanks.

You can close the SSH session. Tmux continues to run in the background and can be invoked in a future session with tmux attach. It'll be just like you left it, but further along.

 

[J-NAS]

Why not use tmux instead of trying to hack your way around?

Thank you for (re)suggesting this. Once you figure out how to create new windows (Control+B, then 'c') you can toggle between them with Control+B, then 'w'. Super simple. It's awesome being able to run badblocks in parallel.

 

[J-NAS]

All my drives checked out OK. No reallocated sectors. Thanks for the help. :) I'm moving on to setting up CIF shares/users and hitting a brick wall, hard. Will do additional reading before posting a new thread begging for enlightenment...

 

[Ericloewe]

All my drives checked out OK. No reallocated sectors. Thanks for the help. :) I'm moving on to setting up CIF shares/users and hitting a brick wall, hard. Will do additional reading before posting a new thread begging for enlightenment...

I found that carefully reading the manual's basic setup was a good start with the whole permissions thing.

 

[J-NAS]

Yes, the manual is a good start. The trouble is following guides in an attempt to not have to educate oneself fully on a very deep and involving topic, when truthfully my hope is to "set it, and forget it". Once it works, I'm not about to try messing with a good thing. I just thought getting it to the "working" stage would be easier...:(

I was following this guide, and even though it was written in May of this year, it turns out running

Code:

cd /mnt/share
chown root:wheel .
chmod 1777 .

will completely hose your ACL permissions. Didn't check the errata guide, did I? Nope. I sure didn't. Didn't know what an errata guide was. Always learning, I am.

I'm beginning to really dislike the Ars Technica article that made FreeNAS seem more "user friendly" than it is. Or has been to me I suppose. It's blowing my mind that I can't get a damned share to work with write access.

I've had to learn how to nuke my existing samba4 and recreate it, but the trouble is I don't know enough to know whether I've done that properly or not! Things aren't working as expected, but I don't know if that's simply user error, or if I'm running into a number of issues that seem to line up perfectly with reported bugs...

Sigh. I'm going to reload FreeNAS onto another USB key and start again to rule out my bumbling as the culprit. I'll re-read the manual regarding permissions.

Any other timely, still accurate for FreeNAS 9.2.1.8=Release, hold-my-hand-like-a-frightened-little-girl resources anyone can recommend?

 

[Ericloewe]

Yes, the manual is a good start. The trouble is following guides in an attempt to not have to educate oneself fully on a very deep and involving topic, when truthfully my hope is to "set it, and forget it". Once it works, I'm not about to try messing with a good thing. I just thought getting it to the "working" stage would be easier...:(

I was following this guide, and even though it was written in May of this year, it turns out running

Code:

cd /mnt/share
chown root:wheel .
chmod 1777 .

will completely hose your ACL permissions. Didn't check the errata guide, did I? Nope. I sure didn't. Didn't know what an errata guide was. Always learning, I am.

I'm beginning to really dislike the Ars Technica article that made FreeNAS seem more "user friendly" than it is. Or has been to me I suppose. It's blowing my mind that I can't get a damned share to work with write access.

I've had to learn how to nuke my existing samba4 and recreate it, but the trouble is I don't know enough to know whether I've done that properly or not! Things aren't working as expected, but I don't know if that's simply user error, or if I'm running into a number of issues that seem to line up perfectly with reported bugs...

Sigh. I'm going to reload FreeNAS onto another USB key and start again to rule out my bumbling as the culprit. I'll re-read the manual regarding permissions.

Any other timely, still accurate for FreeNAS 9.2.1.8=Release, hold-my-hand-like-a-frightened-little-girl resources anyone can recommend?

If you're using CIFS, set the ownership, set it to Windows ACLs and do nothing more in FreeNAS. Do everything else on a (Windows) Client.

 

[J-NAS]

If you're using CIFS, set the ownership, set it to Windows ACLs and do nothing more in FreeNAS. Do everything else on a (Windows) Client.

Would it be that simple? My setup includes 3 PCs, 2 Macs, 2 linux boxes and a host of DNLA devices*. Trying my best to plan out how to divvy things up with appropriate datasets, and the accompanying permissions. I need two admin users, and a bunch of guests.

I will read the longest documentation on earth if pointed in the right direction. What I'm concerned about is filling my head with information that is now no longer valid (chmod on ACL for instance) and having to un-learn it. A lot of guides were written prior to 9.2.1.8, and it seems there are a lot of changes that are only referenced in the errata.

I'll keep reading--thanks for the tips. :)

*I'm drawn to CIFS for the cross-platform ability and support for ACL. Nice to be able to support previous versions (via snapshots) directly within the host OS.

 

[J-NAS]

Didn't know where to stick this; didn't think it deserved its own thread but wanted to share for the other beginners out there:

Badblocks works. I had a Dlink DNS-320 and DNS-323 complain that two drives failed SMART. That's all they told me. I pulled the drives, dropped them in an ancient box with a Live Ubuntu CD and pulled all the SMART info. One drive had 1 bad block and the other had 38. I was prepared to lose the one with 38, but thought the one with 1 would be on the fence.

Since I was likely going to have to rebuild those NASes anyway, I decided to test the drives that passed as well.

Ran a full sweep of badblocks on all four drives, piping the output to .txt files.

Several days later, the two drives that were on the fence had generated .txt files of 3.3GB and 5GB respectively. Badblocks halted. The other two drives completed with no errors.

I attempted to pull new SMART info from the two bad drives but they were dead.

Like, no longer show up during POST dead. Like, changed them over to different SATA ports and didn't come back dead. Like, didn't show up in an ASMEDIA PCI-E SATA card either dead.

So as a beginner I would confirm the often repeated sentiments of this board; it is well worth torture-testing your drives like this before putting them into production! Those drives survived multiple passes of SMART long tests without incrementing their counters. Trust be told I skipped a DD test, which may well have killed them too--but badblocks did them in.

Only trouble is now I can't zero them before they go to the eco-station. Not that it will likely matter. ;)

 

[Fraoch]

Only trouble is now I can't zero them before they go to the eco-station. Not that it will likely matter. ;)

There's a tool you can use in such a situation.:D

 

[Darren Myers]

Only trouble is now I can't zero them before they go to the eco-station. Not that it will likely matter. ;)

I have a funner approach, do you own firearms?
If yes, skip step 2.
Step 1. Get harddrives,get in car, drive to local gun range
Step 2. rent firearms at local range
Step 3. Have at it
Step 4. Go home
Step 5. Explain why you smell like gun powder to your other half, post photos online
Step 6. Dont worry about the data being recoverable

I make sure my data is unrecoverable http://imgur.com/a/bKL0t

 

[danb35]

I've used the firearms route myself (using a .30 cal rifle--makes fairly spectacular holes through the platters, rather than massive dents in them), and it's pretty fun. But even without something quite as drastic as that or the power drill, badblocks overwrites every block several times, so there shouldn't be any significant risk of your private data being recovered.