February 19, 201115 yr I know this has been posted many of times, but cannot seem to find the answer. I just upgraded my Motherboard and got the Asus M4A785-M. In the Bios it has the settings IDE, RAID, and AHCI. I was not sure which is the preferred setting for unraid. The default setting was IDE, however I changed it to AHCI because I thought I read somewhere on here that was preferred. I am around 30% into Parity check and surprisingly my new setup seems to be going a lot slower then my old setup. The new setup has a better MB, processor, and twice the ram. So I am a little surprised and wasn't sure if it was because I selected AHCI in the Bios. My current rate is 57,929 KB/sec and the last parity check I did with my old setup was 75,891 K/sec. All the drives are the same and the only thing I did different with the new setup was disable disk 4 since I moved all the data to another disk and plan on using that empty slot for preclears on my new drives. So my question is IDE or AHCI preferred? Also, any idea why my Parity Speed check dropped so much after upgrading my parts? Here is my old vs new setup. Old Setup: GA-K8N51GMF-9 Athlon 64 CPU 1GB DDR 400 ram Parity Sync = 75,891 KB/s New Setup: M4A785-M AMD Sempron 140 2GB DDR2 800 ram Parity Sync = 57,929 KB/sec On thing I noticed is the new setup boots a lot faster and within 40 seconds from power on I can access the unraid from my browser. In the past with the old setup it took a around 2 minutes until I could access it. Any help would be much appreciated. Thanks.
February 19, 201115 yr Author Here is a syslog. Is it possible a new parity check is slower then when you already have a good parity and just do another? I only ask because the last check on the old setup was when it already had a good parity check. Now on my new setup I disabled disk 4 and had to use the "Initconfig" command to build a new parity. I doubt it really matters though. I guess when this parity check is done I could change the BIOS to IDE and see if the parity check speed improves at all. Syslog_new.txt
February 19, 201115 yr I do not see the finalized parity sync numbers in your syslog. I suspect you're comparing apples to oranges. Only the finalized parity sync times reporting in the syslog should be compared. You could likely be comparing the first 1% of a parity sync to the end 70% of a parity sync operation. All mechanical drives tend to slow down as you get closer to their end. Another difference seems to be this current parity sync operation is calculating parity, not simply verifying. You are comparing read+write to read-only numbers. Feb 18 07:14:25 Tower kernel: mdcmd (9): start NEW_ARRAY
February 19, 201115 yr I have the same mobo and I have SATA ports set to AHCI. That's the correct setting from what I've read.
February 19, 201115 yr Author I do not see the finalized parity sync numbers in your syslog. I suspect you're comparing apples to oranges. Only the finalized parity sync times reporting in the syslog should be compared. You could likely be comparing the first 1% of a parity sync to the end 70% of a parity sync operation. All mechanical drives tend to slow down as you get closer to their end. Another difference seems to be this current parity sync operation is calculating parity, not simply verifying. You are comparing read+write to read-only numbers. Feb 18 07:14:25 Tower kernel: mdcmd (9): start NEW_ARRAY The parity check is done here are the finalized results: Feb 18 16:51:55 Tower kernel: md: sync done. time=34651sec rate=56376K/sec So that is pretty much in line with the speed it was at around 30% of the way in. It is probably like you said and due to the fact this current parity sync operation is calculating parity, not simply verifying it.
February 20, 201115 yr In addition to AHCI you also need to look for a setting called 'IDE-SATA combined mode' (or something like that) and disable it. That option will force your SATA ports 5 and 6 to operate at IDE speeds, even if the rest of your ports are operating at AHCI speeds.
February 20, 201115 yr All 6 drives are at AHCI. Feb 18 07:08:27 Tower kernel: scsi1 : ahci Feb 18 07:08:27 Tower kernel: scsi2 : ahci Feb 18 07:08:27 Tower kernel: scsi3 : ahci Feb 18 07:08:27 Tower kernel: scsi4 : ahci Feb 18 07:08:27 Tower kernel: scsi5 : ahci Feb 18 07:08:27 Tower kernel: scsi6 : ahci but you do have HPA... ata6.00: HPA detected: current 1953523055, native 1953525168 ata1.00: HPA detected: current 1953523055, native 1953525168
February 20, 201115 yr Author The HPA is the reason I decided to upgrade my motherboard from Gigabyte to Asus. Those HPA's were created a while back and currently I am in the process of running preclear right now on 3 new hard drives and the first thing I am going to do is replace the two disks with HPA so I can upgrade to 4.7
February 20, 201115 yr The HPA is the reason I decided to upgrade my motherboard from Gigabyte to Asus. Those HPA's were created a while back and currently I am in the process of running preclear right now on 3 new hard drives and the first thing I am going to do is replace the two disks with HPA so I can upgrade to 4.7 Remember, replace them one at a time. Let the array re-construct onto the replacement. Then replace the other. If you were to replace both at the same time the array would not start as it would be equivalent to a 2 drive failure.
February 21, 201115 yr Author The HPA is the reason I decided to upgrade my motherboard from Gigabyte to Asus. Those HPA's were created a while back and currently I am in the process of running preclear right now on 3 new hard drives and the first thing I am going to do is replace the two disks with HPA so I can upgrade to 4.7 Remember, replace them one at a time. Let the array re-construct onto the replacement. Then replace the other. If you were to replace both at the same time the array would not start as it would be equivalent to a 2 drive failure. Right now my Disk4 is unassigned since I was able to move all the data onto disk5. Since the new drive is 2TB, I will be able to move all the data from Disk2 and Disk3 onto Disk 4 since those were only 1TB drives. After that is done, I will be able to go back and unassign Disk2 and Disk3, upgrade to unraid 4.7 since my HPA drives are gone. Then I will install the new drives and run preclear on both since they will not be in the array. Then once the preclear is done I should be able to just assign them back and be good with 4.7 and no HPA's. So here is my plan. 1) Run preclear on new drive in disk4 slot 2) Assign disk4 to array and copy disk2 and disk3 data to disk4 (those had HPA issues) 3) Unassign disk2 and disk3 and upgrade to 4.7 since HPA drives are gone 4) Add new 2TB drives to slots for disk2 and disk3 5) Run preclear on disk2 and disk3 6) Add disk2 and disk3 to unraid array So when it's all said an done, hopefully I am running 4.7 with 9TB of space and no HPA issues After going to 4.7 should I run the preclear on disk2 and disk3 with the MBR-4k-aligned? The drives (Hitachi 2TB 5K3000) are not AF, so was not sure if it is necessary.
February 21, 201115 yr Author Preclear was successful with no errors. This was the first time I ever used it and and thanks Joe L. for the script. Total time was 24:31 on a Hitachi 2TB 5K3000 drive. Attached is the syslog. 1st_preclear_results.txt
February 22, 201115 yr Pre-clear does not remove HPA. See here: http://lime-technology.com/forum/index.php?topic=10866.0
February 22, 201115 yr Author Pre-clear does not remove HPA. See here: http://lime-technology.com/forum/index.php?topic=10866.0 I was not expecting preclear to remove the HPA. I have a new MB now and will be upgrading disk2 and disk3 from 1TB to 2TB. So between those two I should be free of HPA.
Archived
This topic is now archived and is closed to further replies.