N5200 still not creating RAID 5 -CONTD.

[thbp]

i am trying the 3rd time to get this to create a RAID 5. 2nd try was taking WAY too long. I don't think it was creating RAID. I'm now using 4 750GB WD7500AAKS drives instead of 5 because for some reason, N5200 could no even detect one of the drives anymore (lcd: 00X00_). no idea what that's about... ok so now here's what i'm doing on 3rd try.

1. reset to factory default.
2. upgrade firmware to 1.0.10
3. when i click button to create RAID, i waited until the raid successfull page came up. before, i waited but closed web interface because it looked like page froze. it took about 5 minutes for the success page to open. i thought this might be the reason my other 2 attempts to RAID had failed.
4. when start RAID, i see one green light for one HD going on and off. now has stopped turning green after about 1.5 hours.
5. right now, none of lights green but all have purple lights on.
6. i looked at CPU Loading and it moves between 0-6%. lots of times, it stays at 0%. is that normal? i'm guessing if its RAIDING, it would have higher CPU loading percentage and not be going to 0%.


like i said before, LCD and web interface says its formatting RAID, but i CANNOT see any percentage complete under Storage > RAID. just blinks saying "Formatting RAID ...". everything else on that page is 'N/A'. what im trying to understand is when N5200 is creating RAID, what should i be seeing that would help me know if its RAIDING or not? should i be seeing all green lights going on and off while RAIDING? should i be seeing high CPU Loading percent? its been creating RAID (hopefully...) for about 3 hours now.

[ikonia]

it looks like your getting no-where on this using the forum. Have you tried contacting thecus.

[thbp]

yeah it should NOT be this hard. i've set up a different NAS before. this was not something i thought i would have problems with. have not tried contacting thecus. many complain they don't reply back. looking at their website, it looks like they try to make it as hard to contact them as possible. looks like one person from thecus replies here but hasn't replied to mine.

[peterfu]

Think You should raise a ticket on the theucs website with an exact description. I assume they will ask You to open then access to Your N5200 for them to take a closer look.
Have You noticed, that Your disks are not listed in the document of supported Harddisks, so may be they don't work on the N5200 ?
br
Peter

PS : CPU Usage should be much higher during RAID building

[thbp]

hey peter,

thanks for feedback. after my first try to RAID, i feared that HD not compatible might have been the reason. only after this problem, did i check the HD compatible list. did not know there was one until i started looking. even on their site, list is a little hard to find... looks like all of the WDs besides the enterprise class drives are compatible. guess not anymore... these AAKS models are very popular and would probably be in Thecus's best interest to support this drive from firmware upgrade if possible. under Storage > Disks - the HD model, capacity, total capaicty, etc. were all listed correctly so i thought this model was ok.

just ordered 3 750GB Seagate 7200.10 ST3750640AS drives to retry RAIDING so i'm hoping this will work once i receive them. i think as long as the % complete shows, it really is RAIDING.

[h0mersimps0n]

After three weeks of struggling with my thecus crashing during RAID 5 building process I finally figure out (ON MY OWN) that it was actually a problem with one of my hard drives.

One of my three identical 500Gig Seagate 7200.10 drives was bad. The kicker is it didn't throw out any SMART errors or anything. If you run the INFO module (I think that's the one I used) IGNORE anything that says ok- simply compare the SMART variables/values kicked out by each drive.

It turns out for me that the drive I suspected was a bad cookie (call it a hunch) was throwing "raw read error rate" #'s that were 2-5x's what the other two drives were producing.

Being that you have so many drives in that thing, you should be able to spot the culprit in no time.

SMART is only as useful as the software you use to evaluate the values. I've discovered each manufacturer has different acceptable ranges so pretty much most of the time all you have to go on is comparing to another drive.

I purchased a replacement drive, stopped trying to use the bad drive and it rebuilt first try.

Hope this helps. I'm going to post the same thing on my thread so hopefully others can benefit from my pain.

[peterfu]

After three weeks of struggling with my thecus crashing during RAID 5 building process I finally figure out (ON MY OWN) that it was actually a problem with one of my hard drives.

One of my three identical 500Gig Seagate 7200.10 drives was bad. The kicker is it didn't throw out any SMART errors or anything. If you run the INFO module (I think that's the one I used) IGNORE anything that says ok- simply compare the SMART variables/values kicked out by each drive.

It turns out for me that the drive I suspected was a bad cookie (call it a hunch) was throwing "raw read error rate" #'s that were 2-5x's what the other two drives were producing.

Being that you have so many drives in that thing, you should be able to spot the culprit in no time.

SMART is only as useful as the software you use to evaluate the values. I've discovered each manufacturer has different acceptable ranges so pretty much most of the time all you have to go on is comparing to another drive.

I purchased a replacement drive, stopped trying to use the bad drive and it rebuilt first try.

Hope this helps. I'm going to post the same thing on my thread so hopefully others can benefit from my pain.

Had the same problem with the "raw read error rate" with one of five disks on a already working RAID5, which was suddenly degraded, but all disk status was OK on the disk status page - the INFO module gave me then the hint and replacing the disk did it.
Problem is that You can't install modules if You have no working raid, cause modules are stored on the raid filesystem.
br
Peter

[thbp]

thanks a LOT for your help guys. really helpful
hoping the new drives coming tomorrow will fix raid problem.

to h0mersimps0n: when your thecus was crashing from raid build, was it showing % complete in web interface? also when it crashing, how did you know it had actually crashed since it seems N5200 does not say much at all as to its current status. by the way, after all your raid headaches, is your N5200 good now? are you happy with it overall?

[h0mersimps0n]

To answer your questions:

1.) Yes, a rebuild % done would show up on the thecus LCD, you could watch the % grow in the web interface. For a while I thought the web interface was crashing it so I would just watch the LCD. I knew it was crashed because the LCD stopped scrolling between IP, etc.

2.) Thecus seems good now, uptime: 1 day, 22 hrs, 20 min no problems so far.

3.) Despite how tempermental the Thecus seems to be with hard drives I can't fault it for the problems I had. I connected the "bad" hard drive to my desktop and ran extensive hard drive utility programs- none of them found a single problem. All I know is anytime that hard drive was involved in reading or writing whether it be in my desktop or the thecus the drive would throw an error after a few seconds of starting a process.

IMO (cause you asked) The N5200 is a GREAT IDEA only semi-well implemented. If the company matures it's customer support and more aggressively updates firmware it can maintain it's crown as one of the highest performing NAS on the market (see Tom's article) for the average home user (which is who their demographic should be).

[thbp]

after installing new seagates and starting the raid build, i can tell its really building RAID because

1. under storage > raid, i'm now seeing lots of info as opposed to "N/A" including % complete.
2. also see all 3 HD green lights on showing activity unlike previously.
3. after about 20 minutes, the build % is increasing.

so there you have it. the WD AAKS models (at least the 750GB models) are NOT compatible with N5200. sadly, i doubt i will be the only one to have tried using those drives to build RAID as those are quite popular models. the worst part is, when using those drives, it does seem like its really building RAID from web panel and lcd so its quite misleading. if i had never set up a NAS before, i would have probably been pulling my hair out.


to h0mersimps0n:
2. hopefully our N5200's prove to be reliable. that is after all, one of the strongest reasons someone might want to purchase this.

3. SMART apps not detecting a single problem with a faulty drive is very concerning. but a module could. perhaps that dev should create a SMART program for the desktop.

i had actually read tom's review before purchasing. i came from experience and owning infrant readyNAS. everything you just described about good customer service and updated firmware is like infrant. although with the NETGEAR acquisition, i'm not so sure about customer support anymore... their firmware is pretty mature already though. have you tried their NAS? i was running out of space with readyNAS so i went with this because it had 5 bays and it is affordable for what it is. it also looks like readyNAS is no longer selling diskless models anymore... (bad move IMO). the irony for me is that i had different problems with both NAS. i doubt most NAS are designed for the average home user even though they try to market themselves as that. fact is, most would be beyond confused about Networking, RAID, FTP, Port Forwarding, etc.

[singaporeguy]

hey peter,

thanks for feedback. after my first try to RAID, i feared that HD not compatible might have been the reason. only after this problem, did i check the HD compatible list. did not know there was one until i started looking. even on their site, list is a little hard to find... looks like all of the WDs besides the enterprise class drives are compatible. guess not anymore... these AAKS models are very popular and would probably be in Thecus's best interest to support this drive from firmware upgrade if possible. under Storage > Disks - the HD model, capacity, total capaicty, etc. were all listed correctly so i thought this model was ok.

just ordered 3 750GB Seagate 7200.10 ST3750640AS drives to retry RAIDING so i'm hoping this will work once i receive them. i think as long as the % complete shows, it really is RAIDING.

Interesting I also have 5xWD7500AAKS drives in my Thecus n5200 pro. Initially one of the drives had problem, RAID was degraded overnight. I had the drive replaced with the reseller and so far have no problem with the 5 WD AAKS drives. Running 4+1 RAID5 for the past few days and copied over 400GB of data so far with no errors. Last night I started the process of converting the array from 4+1 Raid5 to 5 disks Raid 5 (currently still in progress). I am running firmware 2.00.00.

[Stefan Huebner]

For this kind of weird errors you should follow those steps:

- install and enable the SYSUSER and SSH modules
- ssh your box
- smartctl -A /dev/sdX (replace X with a-e for the RAID-disks)
- look for following attributes:
Reallocated_Sector_Ct, Seek_Error_Rate, Current_Pending_Sector, Multi_Zone_Error_Rate
... Reallocated secors: if this number is larger than zero, you should replace the disk. It's a prefail value...
... Seek error rate: also a prefail attribute. At first, it slows down operation, at the end your disk breaks...
... current pending: a sector on the disk can not be read. if there's data stored on it, the data probably is lost - if you have a working healthy raid, then you should find out the sector number, erase it's contend, and then issue a rebuild. Easiest way to do that: take out the disk from the working machine, put it back in (now the raid is degraded), reboot and issue raid-rebuild. When those sectors get written to, they will be reallocated and you do not have data-loss. but it takes a good bunch of time and if during this time another disk fails, you loose all your data. So a backup is always recommended... oh, and don't forget to let the reseller replace your disk...
...multizone: also a prefail which slows down at first, then the disk breaks

if everything looks fine, but you still experience weird behaviour, then disable all services and run long selftests on the drives (do it in parallel)
smartctl -t long /dev/sdX
afterwards read out the outcome:
smartctl -l selftest /dev/sdX

When there's still no errors, I wouldn't know the problem... In most cases: the RAM module.

[peterfu]

For this kind of weird errors you should follow those steps:

- install and enable the SYSUSER and SSH modules]
............

This will not work for problems with raids which are not created, cause for the module install You first need a working raid
br
Peter

[Stefan Huebner]

HMMMPFFF

yeah, you're right...
...then let's change it.

try to build a JBOD with only one disk. then follow the steps from my last post.

Oh, and if this also fails, try again with another disk (so you got up to 5 tries alltogether, one should work).