Data Rebuild Extremely Slow

[LRFalk01]

unRaid 4.6

2Ghz single core CPU

~1GB of ram

 

New Drive: WDC_WD15EARS (jumpered to disabled advanced format)

 

Syslog attached.

 

Data rebuilds in the past have gone ~20MB/sec on this machine. With this new drive the data rebuild is going at ~2MB/sec which is going to take over a week to finish. It has to be something with this drive, but I am not sure what to do as far as troubleshooting.

 

Thanks,

Lucas

syslog-2011-04-13.txt

[dgaschk]

The syslog doesn't show enough. Please post another.

[LRFalk01]

I pulled an updated syslog, but it has not changed since the one I posted an hour ago.

 

Edit:

I stopped the data rebuild. It was getting to the point where it would take over 4 weeks to complete. I started a preclear on the drive, and it was going just as slow (copying 0's to the drive @<1MB/sec). Tonight after work I am going to try a preclear on a different computer to see if it is the same. Also going to run some benchmarks on it.

[Joe L.]

I pulled an updated syslog, but it has not changed since the one I posted an hour ago.

 

Edit:

I stopped the data rebuild. It was getting to the point where it would take over 4 weeks to complete. I started a preclear on the drive, and it was going just as slow (copying 0's to the drive @<1MB/sec). Tonight after work I am going to try a preclear on a different computer to see if it is the same. Also going to run some benchmarks on it.

It could be rebuilding slowly if another disk was being read slowly.  It might not be the new disk.

[LRFalk01]

I would be very skeptical that a current disk is the issue. I did a data rebuild ~10 days ago that went just fine. The only difference is this WD drive. Also with the preclear going at under a meg a second, I think it further points to the drive being an issue.

 

Thank you for your response.

[Joe L.]

I would be very skeptical that a current disk is the issue. I did a data rebuild ~10 days ago that went just fine. The only difference is this WD drive. Also with the preclear going at under a meg a second, I think it further points to the drive being an issue.

 

Thank you for your response.

true, or the port it is connected to is in PIO mode, or the cable used, or the total load on the power supply being too much  and the voltage dropping to where one of the drives is having difficulty.

 

You have the right approach though, test it in another box.  under 1MB/s is not normal unless the drive is in a PIO mode.

[LRFalk01]

I did some googling of PIO mode to figure out what you were talking about. I do not see it in my dmesg as it was here:

 

Mobo is Asus K8N-E with NFORCE3-250 chipset.

 

When I boot the stock kernel, hard drives on the PATA IDE ports are UDMA 6 and all is good.  dmesg shows:

 

Uniform Multi-Platform E-IDE driver Revision: 7.00alpha2
ide: Assuming 33MHz system bus speed for PIO modes; override with idebus=xx
NFORCE3-250: IDE controller (0x10de:0x00e5 rev 0xa2) at  PCI slot 0000:00:08.0
NFORCE3-250: not 100% native mode: will probe irqs later
NFORCE3-250: BIOS didn't set cable bits correctly. Enabling workaround.
NFORCE3-250: BIOS didn't set cable bits correctly. Enabling workaround.
NFORCE3-250: 0000:00:08.0 (rev a2) UDMA133 controller
   ide0: BM-DMA at 0xffa0-0xffa7, BIOS settings: hda:DMA, hdb:DMA
   ide1: BM-DMA at 0xffa8-0xffaf, BIOS settings: hdc:DMA, hdd:DMA
Probing IDE interface ide0.

 

After changing the appropriate (or what I thought were appropriate) IDE drivers to built-in instead of modules to eliminate the initramfs, something's amiss.  Now when I boot my new kernel, the disks are in PIO mode.  Apparently something other than the NFORCE3-250 driver is snagging the ide0 and ide1 ports before the NFORCE3-250 driver:

 

Uniform Multi-Platform E-IDE driver Revision: 7.00alpha2
ide: Assuming 33MHz system bus speed for PIO modes; override with idebus=xx
Probing IDE interface ide0...
hdb: JLMS XJ-HD166S, ATAPI CD/DVD-ROM drive
hda: WDC WD2500JB-00EVA0, ATA DISK drive
...
...
...
NFORCE3-250: port 0x01f0 already claimed by ide0
NFORCE3-250: port 0x0170 already claimed by ide1

Anyone know how to make sure the NFORCE3-250 driver grabs the ports first?

 

I've never heard of PIO mode before. Thanks for the knowledge ; )

dmsg.txt