[SOLVED] Stupid me, I think I lost the majority of my drives on my array!!

[S80_UK]

I once had a power splitter with the Molex disk drive plugs that had the colours in the wrong order.  Fortunately they were wrong at both ends so it worked correctly, but I ended up swapping them all around anyway just to preserve my sanity.

[garycase]

I once had a power splitter with the Molex disk drive plugs that had the colours in the wrong order.  Fortunately they were wrong at both ends so it worked correctly, but I ended up swapping them all around anyway just to preserve my sanity.

 

Fortunately, electrons don't know what color the wire is they're traveling on    ... so as long as the proper pins are connected, the color is irrelevant.  HOWEVER ... it's very nice for us humans when everything uses the standard wire colors 

 

 

rwickra => Any luck with Donor Drives?  Definitely interested in some feedback on these guys ... sounds like an excellent resource for drives that are electrically failed but mechanically okay.

 

[rwickra]

Garycase,

 

  So here are the updates:

 

1. Norco folks were wonderful and gave me a new set of backplanes. Everything is working great.

2. I bought a Seasonic G-750 Modular PSU (750W) to replace my Corsair HX750 PSU. However, both PSUs check out normally when I tested it with a PSU tester, so perhaps the Corsair one isn't bad after all. But I am paranoid. I've heard really good reviews about Seasonic. The problem I have is that the Seasonic G-750 has only 4x SATA/peripheral modules, and I have 5 backplanes I was planning on supplying each with a separate cable from the PSU. If I use this, I will have to split 1x modular cable to two backplanes. That means that that PSU cable will be carrying the current to supply 8x hard drives (assuming I am running all 20 hard drives on my Norco 4220 at peak capacity). Do you think this is wise? Or should I just take the risk and use the Corsair which has 6x SATA/peripheral modules and will easily enable me to run 1x modular cable per backplane. What do you think? (I will post this question separately on the forum).

 

3. And lastly, about Donor Drives. I have so far, have had a great interaction with them. I shipped my 4x drives to them that had burned out. They were very responsive, kept me updated via emails and just two days ago notified me that the PCB swap/adaptation process was successful for 3 of the 4 drives. One of the drives they said had a drive-head failure and will require opening in a clean-room if I need the data recovered. I figured that since I had the parity drive intact, that hopefully I should be able to rebuild the data on the missing drive once I get the other drives back. They shipped the drives back to me via FedEx 2day (cost me $16). The total cost of PCB swap/adaptation/diagnostics cost me a grand total of $180. All in all, I would say it was very positive. I am yet to receive the drives (I expect to get them tomorrow so I will try them out), start the array and keep my fingers crossed that the data can be rebuilt and I can still recover the data that I lost.

 

[garycase]

Excellent news r.e. the drives.

 

If you have done NOTHING on the array since this happened, you should indeed be able to recover the last drive via a rebuild.

 

One potential problem:  When you next start the system, it may start an automatic parity check (if it wasn't shut down correctly the last time it was used).    You do NOT want this to happen -- as once it starts modifying anything on the parity drive, you can't do a successful rebuild.  Hopefully that's not the case -- it sounds like you had in fact shut the system down normally, and then had this issue ... in which case it should boot without that.  But be on the alert ... and IMMEDIATELY abort the check if one starts.

 

As long as that's not an issue, you should be able to just start the array with the bad drive not connected -- it will show as "missing";  then shut down; put in a new drive; boot; assign the new drive to replace the missing drive; and Start the array again.  It should rebuild ... and a few hours later you'll have all your data back !! 

 

[Fireball3]

has only 4x SATA/peripheral modules

It doesn't matter as long as the PSU has a single rail.

I'm using an HX650 and have 11 drives connected to one PSU-plug.

To connect the next 10 drives I might use a second one though.

 

But be on the alert ... and IMMEDIATELY abort the check if one starts.

For these special cases, it would be useful to have a setting - be it in a config file or be it a safe/alt mode

of unRAID - so that one can boot without automatically starting the array.

 

Edit:

A timer would also be a solution to give some time to cancel the automatic parity check.

Can we suggest that to Tom?

 

 

[172pilot]

I once had a power splitter with the Molex disk drive plugs that had the colours in the wrong order.  Fortunately they were wrong at both ends so it worked correctly, but I ended up swapping them all around anyway just to preserve my sanity.

 

I've had one of those too.. but I've also had a molex "Y" cable where the pins were actually wired wrong on one of the plugs.  I always check now before using any "Y" or extension cable..

 

I've also had great luck swapping boards, although my experience doing so was in 1992, when my ONE GIG SCSI drive was $1000 and state of the art, and I managed to kill the board with static by putting it down on carpet..  Easy swap in those days.. Sounds like it's a bit more involved (firmware, etc..) today, but still worth a shot!

[DaleWilliams]

I've also had great luck swapping boards, although my experience doing so was in 1992, when my ONE GIG SCSI drive was $1000 and state of the art, and I managed to kill the board with static by putting it down on carpet.. 

I still have a static mat on my desk! I'd just as soon NOT know how many times its saved my data! 

 

 

EDIT: CORRECTION: It's an ANTI Static Mat!   

[garycase]

I've also had great luck swapping boards, although my experience doing so was in 1992, when my ONE GIG SCSI drive was $1000 and state of the art, and I managed to kill the board with static by putting it down on carpet.. 

 

1GB for only $1000 ??  That's 180 times LESS expensive /MB than my first hard drive (1981) ... a 26MB Seagate 14" unit that cost me $4500 (which was a 10% discount off the $4999 list price) 

[ijuarez]

 

I've also had great luck swapping boards, although my experience doing so was in 1992, when my ONE GIG SCSI drive was $1000 and state of the art, and I managed to kill the board with static by putting it down on carpet.. 

 

1GB for only $1000 ??  That's 180 times LESS expensive /MB than my first hard drive (1981) ... a 26MB Seagate 14" unit that cost me $4500 (which was a 10% discount off the $4999 list price) 

 

And they said that would never run out of space

 

 

Sent from my iPhone using Tapatalk

[garycase]

When I bought it I certainly felt like it would be enough space to last a L...O...N...G  time 

 

... and it was quite good for a few years.    But clearly time (and disk capacity) has marched onwards far beyond those relatively early days (I've had PC's at home since 1975 ... and remember when a 180KB floppy felt like a LOT of space !!).

 

At the price/MB I paid for that drive in 1981, one of today's 4TB drives would cost $692,307,692 !!!      Anyone who thinks 4TB drives cost a lot today clearly hasn't followed just how much storage costs have DROPPED over the years