New N7700 - iSCSI Setup Question

[jnheinz]

Hi,

I just recently bought and setup a N7700 with 7x1TB WD Caviar Black HDDs. I setup the RAID5 array with only 1% user data.. It did its thing for about 6 hours, then I came back and setup the iSCSI Target Setup.. Made it use the other 99% .. I followed the instructions in the User Manual.. All seemed to come out OK and I had a 5.5TB iSCSI target (volume).

I have a Windows 2003 Server, I installed the latest Microsoft iSCSI initiator... and went through the setup process in the manual and went into Computer Management.. Bam, there it was. 5.5TB. I initialized and it is showing 2 separate pieces of disk under that Disk .. one is about 2TB, the other 3.5TB .. What did I do wrong or what simple thing am I missing? I would like one big 5.5TB partition under my iSCSI disk. I've tried right-clicking on each of them, but it won't let me delete them to make them into one big one.

Thanks in advance.

-Jordan

[dpankros]

There are many variables here. For one, is that 2003 server SP-1? These documents are a good place to start:
http://www.microsoft.com/whdc/device...e/LUN_SP1.mspx
http://technet.microsoft.com/en-us/l.../cc773268.aspx

[jnheinz]

I may have stumbled upon something.. Is the Microsoft iSCSI Initiator capable of using iSCSI volumes > 2TB? I read a post from a year ago stating that it doesnt not, and that they had to use the Starport iSCSI Initiator.

Let me know if I'm heading the right way with this. I'm not currently home, but I am going to give this Starport iSCSI Initiator a shot when I can RDC into that server.

[msx8280]

I have succesfully setup an iscsi > 2TB and is working fine (zfs, raid6, winxp pro that is)

[dpankros]

I think how it is formatted is the big question. If you format it ntfs, you'll see a 2TB limit on some OSes, including some windows server 2003 releases. The info is in those links I posted earlier in the thread.

With zfs, as the previous poster used, there isn't a 2TB limit.

[jnheinz]

Just figured I'd end this thread with my resolution.. Turns out I've never had to deal with GPT partition type.. I simply right clicked on the disk and changed the partition type from MBR to GPT .. it was fine from there.. it recognized all 5.5TB in the one partition.

One other question.. My N7700 went on the fritz today.. I submitted a ticket. I *BELIEVE* the culprit.. Drive #7 is now showing as 54 reallocated sectors... It was fine a week ago or so.

What happened.. I was using it.. Noticed some slowdowns the night prior.. Was transferring 150GB to it over iSCSI.. The PC I was using locked up from my remote desktop point of view.. Had my girlfriend turn it off and back on.. Found out my N7700 was not responsive to ping.. Reset it, it came back up with drives 5 and 7 missing !! I had her pull out drives 5 and 7 and put them back in.. 5 came up fine, but 7 had all sorts of read errors and stuff. The RAID volume was still missing.. so I updated my firmware to the latest, rebooted, and bam.. my RAID volume is now back, but drive #7 is in WARNING mode.. 54 reallocated sectors still.

Do drives that have a growing # of reallocated sectors possess the capability to lock up the entire RAID array and the N7700? I have ordered another drive and will be RMA'ing this one when I get the other. I figured I should have a spare handy anyways.

Can anyone explain what happened here? I've had the N7700 up and running for about 3-4 weeks now.

[jnheinz]

OK -- Update to this.. I wrote all 150gigs no problem.. Then I tried to unRAR something simple.. went REALLY slow and the N7700 CPU was spiking to 100% at times.

Then I tried to burn a DVD image at 2x .. 5MB/s or so.. and the N7700's CPU was spiking to 100% and my program was having problems keeping a buffer from the iSCSI volume. It looked like it was having problems at a certain part.. it was able to pass by it and then it was fine... This exact same thing happened yesterday.

Is this all because of drive #7's newly found "reallocated sector count of 54" .. ?? Or does all of this warrant another explanation?

Should I take it easy on the volume until I get this replacement drive? Should I just pull out drive #7 ?

[icke]

jordan,

what did you do in the end and is it all working now?

cheers,
olaf

[philhu]

It's not supposed to do this, but I had a Raid-1 do the same thing, not even a Thecus.

If a drive starts to fail, but still reads successfully, then sometimes a raid unit will not take the failing drive offline, BUT it will wait for it to finish I/O, forever, so things slow down dramatically or seem to hang

To solve this, I pulled the bad drive, so raids are NOT foolproof.

[jnheinz]

Solution thus far.. still a week in testing.. but seems to be well.... As the previous poster had noted .. Replacing the drive seemed to make the problem go away.... It seems the drive was flaking out, but not enough to make the RAID take it offline.

New drive made the issue go away... Will keep this thread posted if I find out more. I know how useful these threads can be when you are desparately searching for a solution.