Two failing drives -Help!

[childofthewired]

Hi guys,

I have two failing drives in my greyhole pool. What is the best way to replace these while getting the data off of them? Should I boot to a cloning environment and clone the data over, or is there a way to use greyhole to replace the drives live there was in WHS?

I have no cloning in my pool yet.

[spline]

Hi,

it all depends on how much data you have on these drives, how much space you have on the non-failing drives left, how many drives you have overall in your hda, and if you have a spare drive where you can copy the files on.

There is more than one solution for this.

You can tell greyhole that one drive will go out of the pool. Greyhole will than copy the data from this drive to other drives and you can replace this drive.
You have to do this twice. But i would not recommend this, because it could be, that greyhole will copy some data to the other failing drive.

If you have some more drives with enough space in your hda you can tell greyhole alsways use max copies of the shares. Greyhole will then copy all files to all drives (if there is enough space). So you could simply put out thes drives from the pool as all files should be somewhere on other drives. After replacing the failing drives, you could reduce the number of copies again.

But i think, it's not recommended to give a heavy load on these drives. So i would handle them with care and would do as less operations on them as possible.
What i would do: Have a look what files / shares are located on the failing drives. SSH on your HDA, change to /var/hda/files/drives/failingdrive and use ls to see what greyhole has put on them.
Mount your shares onto another PC over the network, attach an HD onto this PC and copy these files to this.
Aftrewards, tell GH that this drive will go, let GH make it's work (copy the files to somewhere else) and replace it.

But it really depends on the amount of data, free space, and drives left.

Allways remember: Greyhole is no backup solution!!

Good luck


Juergen

p.s. Why are you so sure that the drive is failing?

[sgtfoo]

show us your greyhole drive config... how many drives and how much space is left and how much do you need to keep your data?

You will need to do a drive removal from greyhole and then offload from the /etc/fstab afterwards.

[sgtfoo]

After following spline's steps, don't forget to remove the failed drives from the /etc/fstab using:

Code: Select all

sudo nano /etc/fstab

and get rid of the lines that mount the failed drives, otherwise the next reboot will hang on the fstab due to missing volumes.

[childofthewired]

I'm getting kernel device errors on the two drives, smart test shows them to have escalating bad sector re-allocations. they are WD EARS drives.

Both are getting pretty full, so I think that I could use greyhole to move the data off, once I get a new drive in there. I have somewhere around 7 TB in the pool, with 6.5TB used. I was going to expand the pool with more drive soon. This has just accelerated that.

[ciscoh]

they are WD EARS drives.

thinks back to the lethal weapon movie where Mel Gibson spits out gasoline and mumbles something about Exxon..........i feel that way whenever i see WD.....

i will never own another WD drive again. WD is the worst in my experience, followed closely by Seagate.

just my worthless $.02 rant.

[childofthewired]

What style and brand of drives would you recommend at this point? It seems like so many of them are advanced format now.

[spline]

Hi,

it's funny: I stoll have a 600 MegaByte and 80 GByte WD drive. Both run still perfect. But in the last years WD ruins their name complety. I also own 5 EARS dor EADS drives between 1TB and 2 TB. Almost all of them has this strange "load cycle count" error.
For my HDA i now replaced the system drive with a Hitachi. On my workstation i have a Samsung as system drive.
The WD drives work in my HDA in the storage pool. I power them down when not in use to avoid the load cyce count to increase so fast. And i do regular checks on them. To be honest: I expect them to fail sooner or later and as soon as drive prices will go down again, i will replace them with Hitachi drives.

As you don't have that much free space left in you pool, you should replace them as fast as possible.
To get more free space, you could also try to set the number of copies for your shares to zero (be aware, that this is dangerous as you don's have a "backup" of the files). Depending on the number of copies you had before, this could give you some space back and you could tell greyhole that one drive will leave the pool.

But to be honest: Try to copy the files to an external drive or another PC and turn your HDA off till you have a new drive in place.
It's only a second between increasing number of bad sectors and a complete fail.

Juergen

[sgtfoo]

to minimize data loss, perhaps avoid all other interaction with the drives by shutting down the server until the new drives are in and offloading data from the bad ones.

I've had odd luck between both WD and Seagate. To be honest it's more or less 50/50 with either one being good or bad so I just variate my purchases... my latest drive is a Seagate 2TB 7200RPM drive that I hope won't give me the errors I encountered years ago with the last Seagate drive that failed on me.

The good thing is, most of the companies will RMA a bad drive without questions.

[spline]

Hi,

i just want to know if you sucefully solved your problem?


Regards


Juergen