PRO 6 poor read performance??

To get a baseline performance - first try testing when directly connected. That way if you have performance issues you immediately rule out network effects and narrow the cause down to your NAS, PC or ethernet cable. Also, don't then just assume the NAS is at fault if you still have an issue - it may be, but there are also many such reports on this forum where updated network drivers for instance have had a major impact on performance

sphardy wrote:
... Also, don't then just assume the NAS is at fault if you still have an issue - it may be, but there are also many such reports on this forum where updated network drivers for instance have had a major impact on performance

Ok. I'm not assuming the NAS is the issue, and have been doing all the tests I can think of to rule it out.

Before I tried direct connect, I tried moving the hard drive to my PC (reformating of course), and mounting as Drive V. Then the Netgear test got 122 MB per second both read and write (the actual performance of the hard drive). Then I deleted the NTFS volume, and put it back into the PRO. I also checked network speeds between the two PCs I used for testing. I got > 900 megabits with jperf , so I don't think this is a network driver or cabling / switch issue.

Direct Connect doesn't make much difference. I got about 60 MB read speed, when I reverted back to my normal switched configuration I got in the mid 50s (somewhat better than the original test). Actual results vary some from run to run, so I am thinking there is no obvious proof of a network issue.

And of course 60 MB is still slower than the write speed (which remains around 80 MB), and much slower than the 100 MB read speeds I was expecting (based on the reviews).

So I am still left wondering if there is an issue with the PRO hardware. Temp 1 reads quite high (58 C), temp 2 is a more comfortable 31 C. The drive reports 39 C. Room temperature is about 24 C. Any suggestions on next steps would be welcome.

temp1 is the chipset temnperature and 58°C is quite ok. temp2 is the cpu and also ok. drives at 39 are also ok.

do i understand this right? you only benchmark with 1 drive? i never tried only 1 drive, but that could be the reason.

with 3 or more disks (normal setup) you will get proper speeds.

do you have jumbo frames enabled? if so try disabling them.

beisser wrote:
temp1 is the chipset temnperature and 58°C is quite ok. temp2 is the cpu and also ok. drives at 39 are also ok. do i understand this right? you only benchmark with 1 drive? i never tried only 1 drive, but that could be the reason. with 3 or more disks (normal setup) you will get proper speeds. do you have jumbo frames enabled? if so try disabling them.

I started with one drive, I am getting more of course, but don't have them yet. It would be useful to know if that is the issue.

Jumbo frames are disabled, I currently have the MTU set to 1492, since that is what my wifi uses.

you could always try viewtopic.php?f=48&t=43747

this addon will run a local disk benchmark on your nas and leaves networking alone.

maybe you can post the output here.

Thx for the link, great plug-in. Is there a similar one that does a network test?

Sequential input (block) 85944 k/s. 10% CPU
Sequential output (block) 114293 k/s. 15% CPU

So the read speed is slower than the write by 30 MB/s.

I've written a ReadyNAS to ReadyNAS NFS test a while back I don't think I wrote it very well but it gets the job done.

I agree with beisser that testing with one disk in the NAS that would be your bottleneck (if the client PC isn't). Testing with all drive bays full would be best. I'd suggest testing with at least three to four disks in the NAS.

Take a look at the performance expectations article at http://home.bott.ca

-forgot to post the logged Bonnie output - DD Result:59.2MB/s. Not sure what load that represents, but it happens to be about the read speed I get when directly connected to the PRO. When the same drive is connected to the PC directly, Iometer gets 122 MB/s. It would be interesting to hear what other PRO 6 users are seeing with Bonnie.

-I haven't ruled out client NIC drivers, though I am getting lower than expected results on more than one PC. Unfortunately 2 of them are corporate PCs, where I don't have the ability to disable the antivirus software. The third might well have a NIC issue.

-Has anyone measured performance improvements for the basic NAS functionality when the memory is upgraded? I'm leaving the 1 GB as is for the time being, but am curious.

Thanks for the tips, I will update this when I get the additional drives.

i have 3gb in my pro.. no improvement unless you run some memory-intensive addons on the nas.

edit: and i get full speed on my pro, but i have 6 disks in the system, not 1 :)

StephenB wrote:
forgot to post the logged Bonnie output - DD Result:59.2MB/s.

As indicated in the article ReadyNAS Performance Expectations that's pretty much what you'd expect from a single hard drive. You need multiple hard disks in both the NAS and the client, or multiple disks in the NAS and e.g. a SSD in the client, to get speeds quicker than that. Of course you could also use multiple disks in the NAS and multiple client machines.

Speeds can be highly variable on a number of factors.

StephenB wrote:
Not sure what load that represents, but it happens to be about the read speed I get when directly connected to the PRO. When the same drive is connected to the PC directly, Iometer gets 122 MB/s.

What size file are you testing with? You should test with a file that is at least twice as large as the amount of RAM in your PC.

StephenB wrote:
It would be interesting to hear what other PRO 6 users are seeing with Bonnie.

I don't have a Pro 6, but you should see speeds multiple times what you're getting when you have more disks installed (and a redundant volume i.e. no resync taking place), particularly when all drive bays are full.