NETGEAR is aware of a growing number of phone and online scams. To learn how to stay safe click here.
Forum Discussion
Solved
Pro Pioneer - Poor performance X-RAID Raid-6 with 6x WD Red 3TB
For a really long time I thought the poor performance was due to the fact that I was running an old firmware and had never done a factory reset since 2009 (it was recommended in the past on the forum by mdgm as well as by the Netgear support staff because in place updates leaves stuff around). I hadn't got around to it as I had no means to backup my 10TB data.
Recently I had taken the time to backup my data to some archive disks and got around to factory reset my NAS and upgrade it to 6.5.1.
Although everything was a breeze (few problems, like the CPU Fan getting 0 rpm and the hassle of creating a RAID6 volume without taking multiple days) I noticed that the speed of the NAS was nothing compared to what it was when I first got the NAS. Resync speed was somewhere between 30mbyte/sec to 60mbyte/sec whereas when I first got the NAS and got it in RAID6 with 6 1TB disks, I got speeds in excess of 120mb/sec (speeds where limited by the disk I was writing to, not the NAS).
Before the factory reset and upgrade I always noticed the NAS having quite some i/o-wait, which might indicate problems with the disk, or just that it is nearly full and it has to do a lot of stuff to write data to free space on the disk. After the reset + upgrade that should not be a problem, but I still saw some (not as much as before) i/o-wait. And using "iostat -x" it seems to orient itself around the last 2 disks (might have something to do with the parity?).
Although the nas contains 3TB WD Red disks, I was not entirely sure about the disks, so I took the time to give every disk individually a quick spin with HD tune to get an indication of how well it performs.
The results were as follows:
As can be seen, each disk individually should be able to handle well over 120mb/sec (the minimum of all disks on average). So I kind of rule out the possibility of a disk being the problem.
However I'm a bit at a loss here, and am reluctant to start moving the backed-up data to the NAS as I'm not really sure it's a-ok.
Any hints or tips?
For future readers;
The performance impact boiled down to the following things:
- The RAID implementation currently works different in comparison to older versions of the NAS which has impact on performance, but delivers more reliability
- The implementation of NFS (and/or NFSv4) works different and by default works in a more reliable way. Using the "async" option will greatly improve the speed of transfers, but will greatly increase the risk of faulty transfers in case of power-failure and such
- Performance is (obviously) impacted when there are operations running (ie: (initial) (re)syncing of volumes, balancing, scrubbing, defragmentation, simultanious transfers, etc.)
I tested diskspeeds within my system to evaluate the performance impact. My conclusion for now is that in a RAID-5 or RAID-6 setup the set won't perform better than an individual disk (which was the case in older versions of ReadyNAS but seems to not be true anymore, probably due to the first point). When using the async option, the performance is equal to the individual disks. Note; I have *not* compared speeds using CIFS/Samba. I have tested one time and it seemed that the speeds were comparable to NFS with the async option turned on.
21 Replies
Replies have been turned off for this discussion
I don't think resync speed is very good measure of performance. Perhaps try Nastester ( http://www.808.dk/?code-csharp-nas-performance )
FWIW, parity blocks are evenly distributed across all the disks.
BTW how much memory do you have? The new systems all have quite a bit more than the stock amount in the legacy x86 NAS.
Probably didnt came across right in the opening post, I explicitly waited untill resyncing was complete and then did write tests to the NAS (primarily because I didnt know if OS6 would handle things differently what might cause slower speeds).
So it is not measured only during the resync (but the numbers seemed to be the same)
I have measured both using real-life data (rsync using random files, and using large (uncompresable) files) both over CIFS (samba) and NFS. As well as synthetic tests using dd on the nas itself (writing /dev/zero to a file, forcing syncs during each write).
Rsync over NFS:
7364 files to consider datafile.dat 2048163840 8% 35.64MB/s 0:09:59
synthetic dd output:
root@NAS:~# dd if=/dev/zero of=/root/testfile bs=10M count=100 oflag=sync 100+0 records in 100+0 records out 1048576000 bytes (1.0 GB) copied, 17.1221 s, 61.2 MB/s
It got 2 units of, I thought 1GB being 2GB in total. However htop says it's only 1GB total, so I guess its 2 dimms of 512MB. Free dictates it has about 450mb used / 200mb cached / 350mb free:
total used free shared buffers cached Mem: 1012920 653800 359120 0 1468 225988 -/+ buffers/cache: 426344 586576 Swap: 2094844 0 2094844
Edit: as it seems, the unit has 2x 1GB but sees only 1 dimm of 1GB. Maybe, I didnt place it back correclty when I cleaned the NAS
Edit2: indeed my 2nd dimm wasn't fitted correctly, it now displays correctly that it has 2GB in totall
A lot's changed since you got your Pro.
We've done a number of firmware updates including things such as switching from using EXT3 to EXT4, we've added 4k sector partition alignment, GPT support, and tweaked different settings.The RAID s at a lower level than the filesystem, but that would have undergone some updates too. We've altered the priority of different things as well which may affect the performance of the initial sync.
The WD RED disks are 5400 RPM and would have different performance to your old 1TB disks.
It's worth noting that the array is redundant throughout the initial sync after a factory reset.I don't think the amount of memory would be a problem for this. If you think it is download the logs and see if any of the swap space is being used.
Thanks for the further explanation mdgm; I get that things changed, and therefore might affect resyncing. I also get that because it is a 5400 RPM disk it probably could (and would) perform less well than my initial Samsung Spinpoint 1TB (7200rpm).
However I find it hard to believe that when the disks individually performe well above 100 mb/sec, that a raid-5 setup would perform (way) worse than each individual disk. I find that a bit hard to believe :) as raid 5 (and 6) were generally considered to perform better. Not as good as a striping system, but better than mirrod setups and at least a bit faster than individual disks.
I still kind of suspect certain drives, but have no other (solid) evidence for them being faulty except that they sound a bit weird when in the NAS. I might try factory resetting again, and starting with 1 disk and gradually go up to multiple disks and see if I can pinpoint at which disk (if any) the problem arises.
In OS4 it was possible to totally re-initialize disks (so without initial sync). Is this possible in OS6 as well? Or do I always have to sync even when there was no data on the disks?
NETGEAR Academy
Boost your skills with the Netgear Academy - Get trained, certified and stay ahead with the latest Netgear technology!
Join Us!
