Home Server Questions - lots of questions....

[Stux]

So the main concern is the max limit? Low limit being 6.5 is cool and all, but I guess if the max limit is on the order of minutes (which I don't know, can't find any data on it) that is what would cause a system hang, correct?



Sent from my jailbroke iPhone using Tapatalk

The problem is when the default is minutes and you can't set it to seconds.

 

[LIGISTX]

You may be overthinking this. If the limit can be set to minutes and you set it to minutes, you'd end up with the same potential for an unresponsive system, but I can't imagine a NAS drive coming from the factory that way. My assumption is that the out-of-the-box value is suitable for the way the drives are intended to be used.

The problem is when the default is minutes and you can't set it to seconds.

Correct. That is what I was saying. The minimum isn't of much consequence, the max is the concern. And I can't seem to find a max value..... I wouldn't set the max to minutes, but I am saying I don't know what the max actually is.


Sent from my jailbroke iPhone using Tapatalk

 

[joeschmuck]

I wouldn't worry about TLER or the other variants just so long as you purchase a drive which has it. NAS drives should all have this so that would be the smart purchase. But not everyone is smart and some are just lucky that they haven't had a real problem with non-NAS drives. For instance the WD Greens are not NAS but they generally work well. But if you're unlucky then you can have issues. It happens.

 

[Stux]

Correct. That is what I was saying. The minimum isn't of much consequence, the max is the concern. And I can't seem to find a max value..... I wouldn't set the max to minutes, but I am saying I don't know what the max actually is.


Sent from my jailbroke iPhone using Tapatalk

No, the minimum is the problem. You see, some manufacturers like to differentiate their products by either disabling TLER, gimping it, or locking it to long durations.

If you can set it to a short duration, it is saved across reboots, and you do it, then you don't have a problem.

(I'm assuming the max and min are the range of possible settings)

 

[LIGISTX]

No, the minimum is the problem. You see, some manufacturers like to differentiate their products by either disabling TLER, gimping it, or locking it to long durations.

If you can set it to a short duration, it is saved across reboots, and you do it, then you don't have a problem.

(I'm assuming the max and min are the range of possible settings)

O, well, then according to the documentation I think its 6.5 second. Wouldn't this be within the range of "perfect"?

 

[rogerh]

No, the minimum is the problem. You see, some manufacturers like to differentiate their products by either disabling TLER, gimping it, or locking it to long durations.

If you can set it to a short duration, it is saved across reboots, and you do it, then you don't have a problem.

(I'm assuming the max and min are the range of possible settings)

Yes, there does seem to be some confusion here. The TLER (other acronyms are available) is by definition a maximum time the drive will wait. If you set the value then the drive will never wait longer than this before returning a failure and letting the system carry on, doing other things or retrying. The range of possible TLER times is the range over which it can be set in setting up before using the disk. You want to be able to set it for popular NAS purposes to be no longer than about 7 seconds, but it is possible for other purposes you many want an even shorter time. But having set it, it is by definition a maximum time which the drive won't exceed. Clearly most of the time the drive will react much quicker than this, the TLER only comes into play when a drive operation fails.

 

[Stux]

O, well, then according to the documentation I think its 6.5 second. Wouldn't this be within the range of "perfect"?

Yes

 

[LIGISTX]

Hey guys, I have this SMART data from my 3 TB's, but, I am mostly lost :/

Can anyone tell me anything from this..?

http://pastebin.com/TqDCPCG0

 

[gpsguy]

It's really hard to read in your Google spreadsheet. We'd prefer to have it posted here in code tags or use something like pastebin.

From a quick glance, it doesn't look like SMART tests, either long or short have ever been run on the disks. I believe you can do them in OMV, but don't ask me how.

Since I can't copy/paste information from your spreadsheet, I'll say that the one with serial number: Z1F1MQ97 has a lot of problems. 13k reallocated sectors and almost 1800 current pending sectors.

I didn't look at all of them and not being able to copy and paste data back into a reply is a hardship.

Hey guys, I have this SMART data from my 3 TB's, but, I am mostly lost

 

[LIGISTX]

It's really hard to read in your Google spreadsheet. We'd prefer to have it posted here in code tags or use something like pastebin.

From a quick glance, it doesn't look like SMART tests, either long or short have ever been run on the disks. I believe you can do them in OMV, but don't ask me how.

Since I can't copy/paste information from your spreadsheet, I'll say that the one with serial number: Z1F1MQ97 has a lot of problems. 13k reallocated sectors and almost 1800 current pending sectors.

I didn't look at all of them and not being able to copy and paste data back into a reply is a hardship.

http://pastebin.com/TqDCPCG0

 

[Mirfster]

You can get a decent idea of what to look for from the S.M.A.R.T Wiki ("Known ATA S.M.A.R.T. attributes" Section).

Agree with gpsguy that none of these disks appear to have ever had any SMART tests ran (Short or Long)

SMART Extended Self-test Log Version: 1 (1 sectors)
No self-tests have been logged. [To run self-tests, use: smartctl -t]

I did not go through the results thoroughly but you should get an idea of what to look for from the wiki.

 

[joeschmuck]

Serial Numbers Z1F1MQ97 and W1F21S49 are very bad, replace it at once. Go look at the hard drive troubleshooting guide here at the forums so you can see what ID values are important to look at. Also, run the Long SMART test on the rest of your drives if you are serious about validating them, it's a simple command and in the guide.

 

[gpsguy]

The OP is still using OpenMediaVault. I don't know if he has SSH access to the FS or not.

I suggested that he try to look at the results now, since he was basing his buying decisions (more 3TB drives) and vdev width/pool layout, thinking there was nothing wrong with his existing drives. To paraphrase "they are only a couple of years old".

... it's a simple command and in the guide.

 

[LIGISTX]

Its with my buddy who is actually the software guy, so while I could SSH in from where I am, he can also ssh in from within his network lol.

That being said. Yes. We know those drives are no Bueno. But, what is the command to run the smart tests? I didn't think they would be the same in OMV, not the same OS. We did find this in the gui, would the long test hopefully be what we want to do?

Sent from my jailbroke iPhone using Tapatalk

 

[Stux]

Yes, the long test is what you want

 

[gpsguy]

I would have done the short test first, then the long test. The short test only takes a couple of minutes, after which you should look at the results. Then run the long test, which will take a couple of hours.

 

[Robert Trevellyan]

short test first

Right, if the short test fails, the extended test is a waste of time.

 

[LIGISTX]

My buddy just finished the short test on all of them. The two bad drives in question (thus far anyways) failed, or errored, he didn't send me anything just said the two we previously identified didn't pass.


Sent from my jailbroke iPhone using Tapatalk

 

[Stux]

My buddy just finished the short test on all of them. The two bad drives in question (thus far anyways) failed, or errored, he didn't send me anything just said the two we previously identified didn't pass.

I would't plan on using a drive which failed the SMART short test in a new build. Might be fine to use as a temporary drive to help with any backup/restore, providing there is sufficient redundancy.

If it failed the test, and is still under warranty, then I guess you could RMA the drive?

 

[LIGISTX]

I would't plan on using a drive which failed the SMART short test in a new build. Might be fine to use as a temporary drive to help with any backup/restore, providing there is sufficient redundancy.

If it failed the test, and is still under warranty, then I guess you could RMA the drive?

Yea we won't be using bad drives in the new build, and I will RMA as many as I can that are still under warranty.


Sent from my jailbroke iPhone using Tapatalk