ReadyNAS 2100 VERY slow

[khauser24]

Hope someone can help me figure this out. Here's what I know:

RAIDiator version: 4.2.19
Performance tab settings: Only enable Disk Write Cache is checked
Network switch model Dell PowerConnect 6248
NIC on your PC: Broadcom NetExtreme 57xx Gigabit
Number of disks and disk models:4 500Gb disks, don't remember the model
The read/write MB/sec numbers from IOMeter or drag & drop test. 8.4MB/s write, 9.3MB/s read (from IoMeter)

Additionally, another NAS on the same switch (different brand) offers performance in the 40MB/s range, and my computer can download from the Internet faster than from the 2100. Soon I will be able to configure the 2100 more aggressively (it isn't on a UPS at the moment, but will be shortly), but I don't think that's going to solve this problem.

The device is installed in a server room, and there are certainly a couple of switches between me and the device, but as noted above my performance through the same switch to another device is much better.

The disks are healthy, there's nothing obvious in the system log.

I will try to get the info on the switch it is connected to if that is required, but for now I'd love some ideas of what I should look at.

Thanks!

[edited 4/23 to include network switch specifications]

[readysecure1985]

Check the network_settings.log to see if you are experiencing dropped packets on the NAS.

To check and see if it's the network itself or the NAS, you can also plug directly into the NAS and perform a large file transfer to see if the speed improves. If it does, then I would be suspecting that something else on the network is causing the issue.

[khauser24]

Thanks for the reply!

Well, there are some on the receive side:
RX packets:292945929 errors:0 dropped:16225083 overruns:0 frame:0
TX packets:158408898 errors:0 dropped:0 overruns:0 carrier:0
To my eye that's 5.5% ... I wouldn't think that'd be enough to cause the performance problem I am having. Also, I don't see that number climb much (I've been watching it...)
-Kevin

[readysecure1985]

That definitely makes my thoughts point to the switch, considering that they are RX and not TX dropped packets.

Have you tried changing the port that it's on? it's not uncommon that one or two ports will be the beginning death of a switch.

Are you teaming? Or using Jumbo Frames?

Have you been able to run the direct connect test?

[StephenB]

It could also be a marginal cable.

Though dropped packets on the RX side of the NAS sounds like the switch might not support 802.3x flow control.

If the switches on the path are managed, you might look at their settings.

[khauser24]

I updated the original post to note the switch type (Dell PowerConnect 6248). I think Stephen B. might have a point about flow control as I can't find any indication on the product specs that it does support it. Still don't see how that would cause this level of degradation though.

Later today I will switch cable and port (maybe switch if another is within reach)...

-Kevin

[edit 4/23 - In the 'Fun to trace the wires' bucket, I originally indicated it was connected to a Cisco Catalyst 2950. Now it is connected to both
Will report on performance after I can reboot device]

[khauser24]

I now have both Ethernet adapters enabled, and going to very different switches (one is a 1Gb, the other 100Mb). Regardless of which I use with IoMeter I get the same results as before. (IE, if I map my drive by IP address to one, run IoMeter, then disconnect and remap to the other IP address, run IoMeter, I get the same results with both (well, similar, very close results), and these are the same as on my initial post.

I suspect the next task will be to set up a direct connect test, but that's going to take some time to arrange.

Any thoughts as to where else I might look?

[mdgm]

Is the NAS configured to auto negotiate the network speed? Does it report a gigabit connection when connected to the gigabit switch?

Are you able to try the new production firmware 4.2.20?: http://www.readynas.com/RAIDiator_x86_4_2_20_Notes

Please go to Status > Health in Frontview, check the model of the disks and click on the SMART+ button and check the firmware revision for each disk too.

[khauser24]

Is the NAS configured to auto negotiate the network speed? Does it report a gigabit connection when connected to the gigabit switch?

Yes and Yes.

Are you able to try the new production firmware 4.2.20?: http://www.readynas.com/RAIDiator_x86_4_2_20_Notes

I could certainly try this.

Please go to Status > Health in Frontview, check the model of the disks and click on the SMART+ button and check the firmware revision for each disk too.

Each of the four disks report: WDC WD5003ABYX-01WERA0 465 GB , 35 C / 95 F , Write-cache ON
They are also all version 01.01S01.

Thanks!

[StephenB]

Please go to Status > Health in Frontview, check the model of the disks and click on the SMART+ button and check the firmware revision for each disk too.

Each of the four disks report: WDC WD5003ABYX-01WERA0 465 GB , 35 C / 95 F , Write-cache ON
They are also all version 01.01S01.

Are you seeing any reallocated sectors, spin retries, ATA errors???

[khauser24]

Please go to Status > Health in Frontview, check the model of the disks and click on the SMART+ button and check the firmware revision for each disk too.

Each of the four disks report: WDC WD5003ABYX-01WERA0 465 GB , 35 C / 95 F , Write-cache ON
They are also all version 01.01S01.

Are you seeing any reallocated sectors, spin retries, ATA errors???

Nope ... looked at the stats while gathering versions. Details:

Drive 1:

Code: Select all

Raw Read Error Rate         0
Spin Up Time                3891
Start Stop Count            20
Reallocated Sector Count    0
Seek Error Rate             0
Power On Hours              4791
Spin Retry Count            0
Calibration Retry Count     0
Power Cycle Count           19
Power-Off Retract Count     9
Load Cycle Count            10
Temperature Celsius         35
Reallocated Event Count     0
Current Pending Sector      0
Offline Uncorrectable       0
UDMA CRC Error Count        0
Multi Zone Error Rate       0
ATA Error Count             0


Drive 2:

Code: Select all

Raw Read Error Rate         0
Spin Up Time                3941
Start Stop Count            20
Reallocated Sector Count    0
Seek Error Rate             0
Power On Hours              4791
Spin Retry Count            0
Calibration Retry Count     0
Power Cycle Count           19
Power-Off Retract Count     9
Load Cycle Count            10
Temperature Celsius         35
Reallocated Event Count     0
Current Pending Sector      0
Offline Uncorrectable       0
UDMA CRC Error Count        0
Multi Zone Error Rate       0
ATA Error Count             0


Drive 3:

Code: Select all

Raw Read Error Rate         0
Spin Up Time                3866
Start Stop Count            20
Reallocated Sector Count    0
Seek Error Rate             0
Power On Hours              4792
Spin Retry Count            0
Calibration Retry Count     0
Power Cycle Count           19
Power-Off Retract Count     9
Load Cycle Count            10
Temperature Celsius         35
Reallocated Event Count     0
Current Pending Sector      0
Offline Uncorrectable       0
UDMA CRC Error Count        0
Multi Zone Error Rate       0
ATA Error Count             0


Drive 4:

Code: Select all

Raw Read Error Rate         0
Spin Up Time                3933
Start Stop Count            20
Reallocated Sector Count    0
Seek Error Rate             0
Power On Hours              4792
Spin Retry Count            0
Calibration Retry Count     0
Power Cycle Count           19
Power-Off Retract Count     9
Load Cycle Count            10
Temperature Celsius         35
Reallocated Event Count     0
Current Pending Sector      0
Offline Uncorrectable       0
UDMA CRC Error Count        0
Multi Zone Error Rate       0
ATA Error Count             0