NETGEAR is aware of a growing number of phone and online scams. To learn how to stay safe click here.
Forum Discussion
bluewomble
Mar 13, 2011Aspirant
Disk Failure Detected...
I've recently purchased a ReadyNAS Ultra 6 along with 6 2 Tb Seagate ST2000DL003 disks (which are on the HCL). I've set up the NAS in a dual redundancy X-RAID2 configuration and have starting copyi...
CitizenPlain
Feb 21, 2012Aspirant
Disregard the post below! It didn't work. I just had a disk fail under these settings.
Summary:
The good: I think I found a fix for this problem.
The bad: It can slow the NAS down significantly.
The (possible) fix: Disable jumbo frames, enable full data journaling.
Does it work: Probably? :D -- Update: No. It doesn't. I just had a disk fail.
Details, history, background:
I wanted to give an update on my latest experiences with this issue. I first posted here on March 27, 2011. After my last incident I posted about, I managed to complete a Time Machine backup and copy all the data I needed to copy to the ReadyNAS Ultra 6 Plus. For 8 months or so, it was simply handling incremental time machine backups from one Macbook Pro, copying the occasional movie or folder of photos, and that was about it. No disk failures. No heavy activity, either. Everything seemed fine.
Two weeks ago, I added a new Mac Pro to the network and attempted to complete a Time Machine backup. Over the course of 5 or 6 days of it simply trying to complete it's first backup of about 220 GB, I had three disk failures. The backup simply didn't complete without the RAID reporting a disk as failed. Each time it was a different disk. Each time, the power was cycled, NAS rebooted, and disk rebuilt in about 12 hours and I'd restart Time Machine.
What I changed:
I wondered if I disabled some features that made the NAS go faster if that would solve the problem. So, under:
System > Peformance > UNchecked "Disable full data journaling."
Network > Interfaces > Ethernet 1 tab > (scroll down) > Performance Settings > UNchecked "Enable jumbo frames."
also, for good measure, I disabled the iTunes Streaming Server, which I realized I was never using. (Services > Streaming Services > iTunes Streaming Server).
That's it. That's all I changed. Left "Disk write cache" enabled as before.
Results:
Good stuff: No disk failures. (yet.) My Time Machine backup completed without a disk failure, which it wasn't doing before. (in fact, it wasn't even completing at all before) I've tried copying a big folder over from one machine (~300 GB, mix of large and small files) while initiating a Time Machine backup on the other machine and had no failures, which was previously prime failure time.
Bad stuff: It seems noticeably slower, though I didn't empirically test it. Watching network IO rate during an AFP file copy of lots of small files and it can be down to 2 or 3 MB/s. Large files (~1GB) will still copy at around 40 to 60 MB/s, though this is still slower than the 80 or 90 MB/s I'd previously get.
Conclusion:
I actually don't care if my ReadyNAS doesn't go crazy fast. Data security / no drive failures is more important to me. I mostly use the NAS as a backup solution for Time Machine, work projects, photos, videos, music, movies, etc. Files get copied there, freeing up space somewhere else, and they sit there for months or years. I don't actually work on projects off the drive (I do video and animation) so the IO speed isn't that big of a deal.
Remaining Questions / Anybody wanna help?:
Because I changed two things, I don't know if it was full data journaling, jumbo frames, or the combination of both that's seemingly fixing the problem. (surely the iTunes server had nothing to do with it.) It would be nice if someone tried just disabling one or the other and seeing if they don't have any more problems. It seems feasible that just doing one or the other thing could solve the problem and the NAS might not take such a performance hit, though I honestly don't have the time or inclination to test it.
If someone wants to try it both ways and can work with the NAS not going at full tilt speed, that would be great. i.e., dear forum reader, go and either turn on data journaling or off jumbo frames and then report back if you have/don't-have a disk failure.
I also didn't do a reliable/repeatable benchmark to check the actual difference in speeds this is causing. Someone might want to take a look at that. Would be nice to know what effect data journaling vs jumbo frames vs both has.
... and, of course, it's completely feasible that I'll have a failure tomorrow and there's absolutely no relation to all the stuff I'm saying. (Update: I just did. It doesn't work.) This just seemed noteworthy enough at the time to mention here.
Looking forward to getting this all sorted out without having to buy an entire batch of new drives. Thanks for reading.
Summary:
The good: I think I found a fix for this problem.
The bad: It can slow the NAS down significantly.
The (possible) fix: Disable jumbo frames, enable full data journaling.
Does it work: Probably? :D -- Update: No. It doesn't. I just had a disk fail.
Details, history, background:
I wanted to give an update on my latest experiences with this issue. I first posted here on March 27, 2011. After my last incident I posted about, I managed to complete a Time Machine backup and copy all the data I needed to copy to the ReadyNAS Ultra 6 Plus. For 8 months or so, it was simply handling incremental time machine backups from one Macbook Pro, copying the occasional movie or folder of photos, and that was about it. No disk failures. No heavy activity, either. Everything seemed fine.
Two weeks ago, I added a new Mac Pro to the network and attempted to complete a Time Machine backup. Over the course of 5 or 6 days of it simply trying to complete it's first backup of about 220 GB, I had three disk failures. The backup simply didn't complete without the RAID reporting a disk as failed. Each time it was a different disk. Each time, the power was cycled, NAS rebooted, and disk rebuilt in about 12 hours and I'd restart Time Machine.
What I changed:
I wondered if I disabled some features that made the NAS go faster if that would solve the problem. So, under:
System > Peformance > UNchecked "Disable full data journaling."
Network > Interfaces > Ethernet 1 tab > (scroll down) > Performance Settings > UNchecked "Enable jumbo frames."
also, for good measure, I disabled the iTunes Streaming Server, which I realized I was never using. (Services > Streaming Services > iTunes Streaming Server).
That's it. That's all I changed. Left "Disk write cache" enabled as before.
Results:
Good stuff: No disk failures. (yet.) My Time Machine backup completed without a disk failure, which it wasn't doing before. (in fact, it wasn't even completing at all before) I've tried copying a big folder over from one machine (~300 GB, mix of large and small files) while initiating a Time Machine backup on the other machine and had no failures, which was previously prime failure time.
Bad stuff: It seems noticeably slower, though I didn't empirically test it. Watching network IO rate during an AFP file copy of lots of small files and it can be down to 2 or 3 MB/s. Large files (~1GB) will still copy at around 40 to 60 MB/s, though this is still slower than the 80 or 90 MB/s I'd previously get.
Conclusion:
I actually don't care if my ReadyNAS doesn't go crazy fast. Data security / no drive failures is more important to me. I mostly use the NAS as a backup solution for Time Machine, work projects, photos, videos, music, movies, etc. Files get copied there, freeing up space somewhere else, and they sit there for months or years. I don't actually work on projects off the drive (I do video and animation) so the IO speed isn't that big of a deal.
Remaining Questions / Anybody wanna help?:
Because I changed two things, I don't know if it was full data journaling, jumbo frames, or the combination of both that's seemingly fixing the problem. (surely the iTunes server had nothing to do with it.) It would be nice if someone tried just disabling one or the other and seeing if they don't have any more problems. It seems feasible that just doing one or the other thing could solve the problem and the NAS might not take such a performance hit, though I honestly don't have the time or inclination to test it.
If someone wants to try it both ways and can work with the NAS not going at full tilt speed, that would be great. i.e., dear forum reader, go and either turn on data journaling or off jumbo frames and then report back if you have/don't-have a disk failure.
I also didn't do a reliable/repeatable benchmark to check the actual difference in speeds this is causing. Someone might want to take a look at that. Would be nice to know what effect data journaling vs jumbo frames vs both has.
... and, of course, it's completely feasible that I'll have a failure tomorrow and there's absolutely no relation to all the stuff I'm saying. (Update: I just did. It doesn't work.) This just seemed noteworthy enough at the time to mention here.
Looking forward to getting this all sorted out without having to buy an entire batch of new drives. Thanks for reading.
Related Content
NETGEAR Academy
Boost your skills with the Netgear Academy - Get trained, certified and stay ahead with the latest Netgear technology!
Join Us!