QNAP: RAID 6 to the Rescue (continued)

Its the onset of Fall and the leaves are starting to make a mound in the backyard. There's been some rain as well. With the seasons changing that also means the room temperature as well.

This morning was the first time I ever saw the advantages of RAID 6.

Actually, the single advantage appears to be stability in the face of two-(2) failed hard drives. For some reason Slot-6 and then Slot-1 started to act up and drop out of the RAID. Now, I know there's nothing wrong with them and my efforts to check the temperature history should little variance. What I did to recover was to simply pull out, dust off, then jammed (literally) the drives back in. This seems to have worked and the RAID just went ahead and rebuilt itself.

Scary, with several NAS once in a while making beeps you get to a point of ignoring these please of alert. Literally the first time Slot-6 fell it wasn't even apparent to me that something was going on as the QNAP is a heavily used file server with gigabytes of transfer going on at all times of the day.

So was it really the weather?

I was going through some forums as well just now and did see some reports similar to what I had just described. Some say firmware on the NAS, some say its the WD20EADS drives -- to me it was the weather and I'll stick to my theory unless something else comes up. :-)

Did this experience help me to reconsider aiming for RAID 5 and recovering 2TB? Maybe, just a little.

However, if things push forward and I get the chance to transfer the main download/upload function to a data center ...? 2TB back for storage please, thank you very much!

QNAP: Add another drive to RAID 6, Part II

Its already Tuesday and the expansion is stuck at 75%. All services don't want to work and the space shown via ssh is only 25-MB, plus /dev/md0 is unmounted. The system is still emitting that irritating double long beep every so often that I hear it in my sleep.

There is too much data on those RAID drives now, now way will I accept it getting lost. Knock on wood! I'm scared and loosing sleep over it :-(

I saw a link to someone having expansion problems in the forum as well as the follow-up solution. In case a reboot won't work there was another way shown. ..... Its a backup plan.

But really, hopefully my process actually works and I can send some more positive feedback to users who would rather not have to mess with the internals and stick to the interface. Though QNAP Tech Support is in California and sound to be very helpful, I'd much rather not have to deal with any form of remote access to my machines as much as possible.

Lets be clear that the SATA Disk view still says that all my drives are healthy. The fact that /dev/md0 is already unmounted may actually mean that at 75% this is in Step #9 of the suggested manual expansion process. :-)

But I'm still not sure :-(

QNAP: Add another drive to RAID 6?

Yesterday I mentioned that everything was working well in terms of building a 7x2TB configuration while the remaining 2TB of data from an external connection was transferred over to the RAID.

Well, this morning all the copying seemed to have finished overnight which meant it was time to plug the last 2TB drive and expand the size of the RAID.

Funny, I was about to say "RAID array" but that is redundant being as RAID does mean "redundant array of individual disks" so saying it would have been like referring to a "redundant array of individual disks ARRAY". No, duh.

It looks to be there is a terminology difference in the jargon used and its confusing. What makes it more difficult is that there isn't enough specific information about what one has to do if using RAID 5 vs RAID 6. After much head-to-knuckle banging (at the possible data loss in one fell swoop of a misconfigured array), a clue on the Wikipedia definition of RAID 6 helped. So it looks to be that its considered by some "RAID 5 + 2 spare disks". Meaning that I could take the "Add hard drive" under the RAID Management tab even if it says only for RAID 5 and implement it on my RAID 6 - which I did. The results are in the first graphic above as well as some intermittent dual-beeping coming from the TS-809 right now.

Those 2 beeps (per the manual, yes of course I read it) mean any of the following things: HDD volume is filling up (don't think so), HDD is indeed full (nope), HDD is in degraded/failed mode (nah), HDD starts/currently rebuidling (YES!!).

So, right now the RAID is rebuilding as a result of the added disk and also calling the expand module. Its still stuck at 1% but like the initial building where I was letting it multi-task copying from the last 2TB drive in the set, its more probably a result of me copying 69-GB of data at 18 MB/s over the gigabit network to one of the NAS' shared drives.

Time to wait and see and a post-update later ....

RAID 6 and that (write intent) Bitmap

For the semi-uninitiated, what is the importance of RAID 6 over RAID 5 is the question. According to 3ware (apparently one of the leading hardware vendors for RAID) say that "Double-parity RAID, commonly known as RAID 6, safeguards against data loss by allowing up to two consecutive drive failures". The technology brief continues to say that RAID 5 is vulnerable to two-(2) types of error conditions: all data is lost in the unlikely event of a second drive failure during a rebuild; and latent media errors that are undetected media defects in the same case as a disk failure on rebuild. Essentially the conditions are the same in that it generally means you have multiple failures in your array instead of the single spare disk you have allocated. There are also performance hit differences, but essentially it looks to be that RAID 6 has a 10% write penalty advantage while RAID 5 has a 15%.

Another good resource introduction to the pro's and con's comes from an article in 2007 saying that RAID 6 is here and viable given how cheaper storage is today. I would have to agree given that USB thumbdrives can now go as large as 64-GB which used to be about the size of my laptop hard drive in 2003. With SSD drives now as large as 250-GB its now viable to get instant startup gratification.

So what about Write Intent Bitmaps? A related discussion on QNAP has some pointers to other external information, but its all not definitive.

Some poor chap basically got flamed on the QNAP forum for asking. Is this the current state of the world today where the smart-asses attack the green apples? Would the preventive thing to do such as a QNAP Korea forum help? Or is this just lashing out payback in some way when non-English based support forums are your only hope for this smashingly great device that you bought from overseas and you can't speak a spit of that other language? Hmmm ... I digress.

To me, given the result of turning on the Bitmap option has alot to do with RAID 6. Thats an over-simplification but too much reading and my best guess (and hope!) is that between the 4-TB spare total I'm going to have on the TS-809 would be better justified if it were also used to ensure that parity and the speed of any possible crash rebuild/resync with write intent bitmap turned on. Given that 2TB initial loss to RAID 5, it had BETTER be.