Since rebuild NAS4FREE is unusably slow

[maraach]

Hi guys,

Been using NAS4FREE for a couple of years now and to date have been very happy with it. Generally average about 100mb/s transfer on large files.

Running a Supermicro X8si6-f board, xeon x3450, 16gb ecc ram and 8 drives on the SAS2 controller in a RAIDZ2 setup.

Now, the problems started when i had a catastrophic hardware failure. Lost 4 drives (bye bye seagate, never again) and the cpu overheated (how did this happen at the same time?). So yes, lost the array... lucky I had a backup.

Now I had a lot of drama's rebuilding everything, had to follow the forum tips on how to "wipe" the old ZFS disks by script as they wouldnt allow me to add them to a new pool. Drives wouldnt come up in format list so had to use the gpart trick. Upgraded to the newest NAS4free distro, wiped my config (as it kept trying to revive the old pool). After 2 days (and very long nights) of googling I finally had a setup which resembled my old one.

Problem. I started copying my data back over cifs / smb.... 30kb/s write speeds.. until about a minute later the connection drops. Yes thats kb not mb. Try other files, other devices, try other nas' i have on the network.. no outcome. Other NAS (which has always been slower) gets a solid 60mb/s. Change routers, network cables check the status 1000t. Try ftping files.... 30kb/sec so it not just cifs. Nothing in log files...

Turned on SMART, checked the drives, all look ok.

Did some more reading wondered if its because i replaced 2 of the drives (previously seagate) with Hitachi NAS drives. Removed them from the pool (now 6 seagate drives) but same problem exists.

Im really at a loss here guys... any suggestions?

[kenZ71]

Is your pool healthy or degraded and trying to rebuild?

Are you using the same memory? Run a memtest to verify the memory?

Try out the tuning tips in the tuning section?

[b0ssman]

go through this list step by step and post the results

http://n4f.siftusystems.com/index.php/2 ... /comments/

[maraach]

iperf results (similar to what im seeing speed wise..)

------------------------------------------------------------
Client connecting to 192.168.1.100, TCP port 5201
TCP window size: 0.06 MByte (default)
------------------------------------------------------------
[184] local 192.168.1.6 port 61214 connected with 192.168.1.100 port 5201
[ ID] Interval Transfer Bandwidth
[184] 0.0- 1.0 sec 0.09 MBytes 0.09 MBytes/sec
[184] 1.0- 2.0 sec 0.02 MBytes 0.02 MBytes/sec
[184] 2.0- 3.0 sec 0.02 MBytes 0.02 MBytes/sec
[184] 3.0- 4.0 sec 0.02 MBytes 0.02 MBytes/sec
[184] 4.0- 5.0 sec 0.02 MBytes 0.02 MBytes/sec
[184] 5.0- 6.0 sec 0.02 MBytes 0.02 MBytes/sec
[184] 6.0- 7.0 sec 0.04 MBytes 0.04 MBytes/sec
[184] 7.0- 8.0 sec 0.02 MBytes 0.02 MBytes/sec
[184] 8.0- 9.0 sec 0.02 MBytes 0.02 MBytes/sec
[184] 9.0-10.0 sec 0.03 MBytes 0.03 MBytes/sec
[184] 0.0-10.9 sec 0.32 MBytes 0.03 MBytes/sec
server sent unexpected data failed
Done.

for dd if=/dev/zero of=/mnt/dataset/test.dd bs=2M count=10000

10000+0 records in
10000+0 records out
20971520000 bytes transferred in 33.066612 secs (634220403 bytes/sec)

for dd if=/mnt/dataset/test.dd of=/dev/null bs=2M count=10000


10000+0 records in
10000+0 records out
20971520000 bytes transferred in 41.827852 secs (501376929 bytes/sec)

No SMART Errors, no errors posting on the pool under information

pool: SeagateRaidZ
state: ONLINE
scan: none requested
config:

NAME STATE READ WRITE CKSUM
SeagateRaidZ ONLINE 0 0 0
raidz2-0 ONLINE 0 0 0
da0 ONLINE 0 0 0
da1 ONLINE 0 0 0
da2 ONLINE 0 0 0
da3 ONLINE 0 0 0
da4 ONLINE 0 0 0
da5 ONLINE 0 0 0

errors: No known data errors

Highest CPU load over 8 cores is 22% during load.

5. Power Daemon - Set to Maximum Performance

6. PCI interfaces - No PCI cards

7. Dont believe my drives support advanced format

8. Tried this command and got "/root: Permission denied."

9. Have not gone down this path yet

10. Pool is currently completely empty

11. DEDUP is not turned on.. i think

12. I do have 2 different types of HDD's but in order to try and resolve this problem i have removed all but 1 brand.

13. Usind RAIDZ2

14. set to AHCI

15. Not using L2ARC

16. - 19. Seem more informational


If im not mistaken item 1 shows that there is a problem? Im not sure how to read the results of 2 but it looks not too bad?

[b0ssman]

yes the problem seems to be with the network equipment.

now you will have to find out where the fault is.

cable, switch, nic.

if you have a third computer see what the performance between those is.

[maraach]

I have tried switching cables (to no avail) and also switched routers. Have numerous other pc's on the same switch and they all operate well... i will have to keep trying.. any more ideas please send them my way! this is killing me ;(

[b0ssman]

have you tried the other network port on the mainboard?

alternatively do you have a network card that you could put in the server for testing?

[maraach]

Ok, with your help i finally found the problem!!

Cant believe it, my virtualisation server had created a bridged connection on my main pc which was causing the problem. I deleted the bridge and we are back to 105mb/s! thank you!!