13 replies to this topic

[dvn]

Came home from work yesterday only to hear the squeal of the RAID alarm going off in my WHS 2011. RDP'd in, opened the WebUI, and saw that the card had failed. My shares were inaccessible. I uninstalled/reinstalled the RAID driver. No good. I rebooted and was surprised to see that RAID init in BIOS looked normal. Back to desktop, problem is gone. Everything's back to normal, like it never happened.

Thoughts?

[ikon]

Those are the kinds of problems that really scare me.

[dvn]

The continuing saga - As you can see, I started a 'verify' operation on the 7th. Apparently, it failed about 3.5 hours into the process. A 'rebuild' started immediately and completed 5.5 hours later. Fifteen hours later, the whole array failed once more. I had shut the alarm off so I didn't notice until now. I rebooted and everything's working once again. I bought what I thought was a good quality product precisely because I wanted to avoid these kinds of situations. This has been a learn as you go process for me, but I'm not liking what I'm learning so far.

Thoughts?

[dvn]

OK. Starting over from scratch. Currently syncing data from WHS 2011 to my EX495. Once done, I'll have a look at each individual drive in the RAID. Assuming all is good, I'll rebuild the RAID 5 and restore the files. Not sure what else I can do at this point.

[tojoski]

Maybe you got a lemon, it happens to even the best hardware mfg's.

Then again, are your drives in a backplane? I had a similar issue when i first set up my raid in the icy dock backplane.

All of a sudden I was getting random disk errors. Turns out it was the molex plug into the backplane just wasn't making good connection.
I could tell because I could wiggle the plug and hear the drives spin up and down. A quick pinch with the pliers and that problem was solved.

I do think the bad sector on channel 6 is suspect. I would pull that drive and so a sector scan with a tool like seatools or spinrite just to be sure.

[dvn]

No backplane in use. Power cables are tight. Breakout cables are well-seated.

Yes, that bad sector. SMART says that none of the drives have reallocated sectors. Maybe 'bad' means that the sector wasn't written properly for some reason other than physical defect? I don't know. I'll test it with something. Not looking forward to this. Testing one 2TB HDD can take some time. Testing all...see you next year. haha

[ikon]

Maybe you got a lemon, it happens to even the best hardware mfg's.

Then again, are your drives in a backplane? I had a similar issue when i first set up my raid in the icy dock backplane.

All of a sudden I was getting random disk errors. Turns out it was the molex plug into the backplane just wasn't making good connection.
I could tell because I could wiggle the plug and hear the drives spin up and down. A quick pinch with the pliers and that problem was solved.

I do think the bad sector on channel 6 is suspect. I would pull that drive and so a sector scan with a tool like seatools or spinrite just to be sure.

I hate those Molex connectors; too tight, too loose, can't get the pins to line up right, bah!

[WebmasterBjerke]

If using RAID controllers from HighPoint tech - rocket raid 2720 http://www.highpoint-tech.com/ (or other raid controllers) and large hard drives, 1 terabyte, 2 TB or 3 TB from Seagate or Western Digital you can experience problems that raid 5 crash if one of the disks just fall out and are offline. There is no smart error tracking. It may be different drives every time and there is nothing wrong with the disks.
What you must do first as a good rule is to upgrade the firmware on hard drives, the raid controller and the latest raid drivers.

The reason it fails is easier than you might have imagined: Interference - radio interference from the cables that go from to controller to the disks are too close to each other.
Try adding some anti-static material between the wires, Split and seperate from each other fully into the connector on the RAID card. There are two ports where up to 8 SATA /SAS drives can be connected on this card, it should have been built so that the 8 ports on the card was located away from each other on the card to prevent noise. There is nothing wrong with the card, more a design flaw, but the cables have to little isolation material and is also tightly packed together these SAS to SATA cables.
PS Graphics, memory modules and other PCI Express cards can also disturbe the data cables to the hard drives. I Recommended that the cables are placed so they do not come in contact with other units! It can also be a layer of anti-static material on both sides of the RAID card to shield it from other devices.
It may fail on some type of PCI Express ports, try to swap the ports if it does not work properly.


If using RAID controllers from HighPoint tech - rocket raid 2720 http://www.highpoint-tech.com/ (or other raid controllers) and large hard drives, 1 terabyte, 2 TB or 3 TB from Seagate or Western Digital you can experience problems that raid 5 crash if one of the disks just fall out and are offline. There is no smart error tracking. It may be different drives every time and there is nothing wrong with the disks.
What you must do first as a good rule is to upgrade the firmware on hard drives, the raid controller and the latest raid drivers.

The reason it fails is easier than you might have imagined: Interference - radio interference from the cables that go from to controller to the disks are too close to each other.
Try adding some anti-static material between the wires, Split and seperate from each other fully into the connector on the RAID card. There are two ports where up to 8 SATA /SAS drives can be connected on this card, it should have been built so that the 8 ports on the card was located away from each other on the card to prevent noise. There is nothing wrong with the card, more a design flaw, but the cables have to little isolation material and is also tightly packed together these SAS to SATA cables.
PS Graphics, memory modules and other PCI Express cards can also disturbe the data cables to the hard drives. I Recommended that the cables are placed so they do not come in contact with other units! It can also be a layer of anti-static material on both sides of the RAID card to shield it from other devices.
It may fail on some type of PCI Express ports, try to swap the ports if it does not work properly.

http://www.webmaster...n.php?side=raid

Edited by WebmasterBjerke, 20 November 2012 - 09:19 PM.

[WebmasterBjerke]

More info:
Also when using 8 disks in a computer make sure you have a strong enough PSU! I recommend minimum 700 watt.

I used one year on making this card work! I used the Highpoint tech support team with it until they ran out of suggestions. I tried tweaking all possible settings in the raid management tool, in bios etc.
I change from Intel to AMD chipset/cpu/motherboard.
I changed ram modules, powersupply( please with 8 disks use a minimum of 700 watt psu) cpu, disk brand from Western Digital to Seagate same problem, raid failed within a day! I also had slow copy and write speed, as slow as 2 mb a second!

(PS Some Gigabyte motherboard refuse to boot from the RocketRaid Controller, stay away from Gigabyte if u use a HighPoint Tech Rocket Raid card, other people on the net say the same thing)

So I think Interference! I remove the external isolation material on the SAS to SATA cables on the side to the raid card. I split each signal cable away from each other, AND NOW IT WORKS without failing.

SO this is the solution. I also added anti static material to the on both sides of the Raid card to shield it from interference!

Now I enjoy read / write speed on SSD disk up to 800 mb a second, and on Mechanical drives (the other raid )is 3 tb x 6 drives in RAID 50, Giving me a total of 12 TB of data. Works fine....Write speed 380 mb a second - read 600 mb a second.

[ikon]

Surely you could just buy better cables. How many different brands/models of cables have you tried?

[Jerry]

I've had a WHS2011 and RocketRaid 2720 with Highpoint cables and four 3TB WD drives running for about a year. Last week the RAID vanished. No alerts or alarms other than WHS reporting the shadow directories could not be found. Logging into the RAID management utility, I couldn't find any errors. Tried to recover files, but all were lost. Noticed there were new bios and driver updates. Updated, initialized on WHS and a quick format and all is working again. Interesting, the RAID broke after I backed up some files the day before and pushed the storage over 3TB. Really spooked this thing could break again. Have a second WHS2011 with a Drobo and four 2TB drives. This one has run flawless for over a year.

[WebmasterBjerke]

Surely you could just buy better cables. How many different brands/models of cables have you tried?

It is mostly standard cables out there, but the SAS to SATA cables seams weaker than a normal SATA cable regarding shielding.
I put more info about raid problem on my page here:

http://www.webmaster...n.php?side=raid
Second part of the page is in English.

[ikon]

I must say I've never had issues like you describe. I do stick with cables 1M long or less. At work I've had large numbers of drives connected to a single card, but they're usually in cages. At home I don't think I've had more than 5 drives connected to a single card. But, I've never had an issue with a RAID setup that I could trace to bad cabling. I've had single SATA cables give me trouble, but not SFF cables. It's interesting to hear your experiences.

[ikon]

BTW, thanks for the great power supply calculator link on your web page: http://www.extreme.o...n.com/PSUEngine