April 19, 200719 yr When I just use my sata drives my parity builds at 35kb/sec... when I add the pata it drops down to 2kb/sec. I have chanted cables 3 times ... Attached is my syslog after reboot with errors.... any suggestions
April 19, 200719 yr I don't know enough to read the logs well, but it sounds like unraid doesn't fully support your pata controller chipset and is using PIO mode instead of DMA (much slower). That was the problem with my board anyway. They might be able to build a custom unRAID kernel for your board. Or use it as an excuse to upgrade the size of your drives, like I did, and get sata ones Then eBay or whatever the old ones.
April 19, 200719 yr One thing interesting is your IDE drive shows up as a 'slave' device - that's not necessarily a problem. How do you have it connected, and is it jumpered for 'cable select'. There are also some other curious messages in the log, what motherboard are you using?
April 20, 200719 yr Author It is the slave and is cable select .. I had two IDE drives connected and disconnected one to see if that solved the issue before I ran this syslog. The motherboard is a Foxconn k8s755a. The manual is here http://www.foxconnchannel.com/service/downloads.aspx?Model=K8S755A-6EKRS&DownID=en-us0000001]ads.aspx?Model=K8S755A-6EKRS&DownID=en-us0000001
April 20, 200719 yr I myself had ran into this bug in the past and switched all my IDE drives to Cable Select and the slow parity build went away.
April 20, 200719 yr All the messages in your syslog that look like this: Apr 18 11:45:23 Tower kernel: [ 92.360209] ata5.00: exception Emask 0x0 SAct 0x0 SErr 0x0 action 0x0 Apr 18 11:45:23 Tower kernel: [ 92.360218] ata5.00: cmd b0/d0:01:00:4f:c2/00:00:00:00:00/00 tag 0 cdb 0x0 data 512 in Apr 18 11:45:23 Tower kernel: [ 92.360220] res 51/04:01:00:4f:c2/00:00:00:00:00/00 Emask 0x1 (device error) Should not be there. Eventually we get this: Apr 18 11:45:24 Tower emhttp[975]: smart ioctl: Input/output error I think one of your SATA drives does not support SMART properly, or is failing, and the I/O subsystem is resetting everything as part of error recovery, leaving your IDE drive in PIO mode. You can try removing your SATA drives one-by-one until the culprit is found by checking syslog and noticing the absense of those messages. Then configure system without that drive and see if faster parity-sync results.
April 20, 200719 yr Author I removed all Sata drives and only connected my pata drives... My parity speed is now on average 1,779 kb/sec I have attached the new syslog w/o the above errors
April 20, 200719 yr After parity-sync has been running a few minutes, then do a 'Refresh' - is the rate still slow? If so, please type these commands and post their output: hdparm -i /dev/hda hdparm -i /dev/hdb
April 23, 200719 yr Made a driver change which might fix this problem for you. It will be in the next release which I'll try to get out today or tomorrow.
April 26, 200719 yr Author Thank you Tom!!! .. My PATA and SATA drives now coexist just fine with parity builds running ~32 - 36 kb/sec. with ver 4 beta 10. I appreciate your prompt response and am now happy I invested in an unraid server. My next question would be what tool or utility can I use in Windows XP to measure the lan bandwidth to my unraid server and the write speeds to the unraid server disks?
April 26, 200719 yr Look around the forum here, I know it has been covered with people posting their results. I just don't wanna search for it
Archived
This topic is now archived and is closed to further replies.