[SOLVED] Preclear problem 3tb

Yesterday morning I started the preclear script v1.13 on my unRAID 5.0-rc5 box.

The drives are WD30EFRX

 

I installed screen and created a new instance in screen (thus leaving me with one instance and two screen windows)

 

command as seen:

screen 1:

./preclear_disk.sg -A -M 4 /dev/sda

 

screem 2:

./preclear_disk.sg -A -M 4 /dev/sdb

 

I checked up on the machine every now and then to see how it was going (between the regular email updates I got)

Everything seem to went fine.

 

Today it was running the Post-Read.

When I reattached the screen I was unable to switch between the screens with C+a "

Nor could I detach with C+a d.

All I could see was the preclear script screen stuck at "Post-Read in progress: 69% complete."

 

I switched to another tty and logged in, tried to reattach screen from there. Did not work.

I switched to a third tty and checked the running processes. It looked like there was very little activity (however I have nothing to compare to since this is my first unRAID and preclear).

 

I checked the time interval of the mails and figured it would take about 3-4h for 25% to be read. So after that I started to worry.

 

I checked the /var/log/syslog

and the only thing I could find (of interest) was the following:

 

Oct 6 17:10:15 omicron emhttp:WDC_.... serial#... (sda) 2930266584

Oct 6 17:10:15 omicron emhttp:WDC_.... serial#... (sdb) 2930266584

which I am guessing only tells me that it successfully found the harddrives and the ammount of space in byte.

An hour later I get this:

 

Oct 6 18:07:01 omicron kernel: sda: unknown partition table

Oct 6 18:07:24 omicron kernel: sdb: unknown partition table

Which does not surprise me since the drives are not partitioned.

 

After that I get the following:

Oct 7 11:00:27 omicron kernel: sdb: sdb1

Oct 7 11:29:47 omicron kernel: sda: sda1

Oct 8 00:11:46 omicron kernel: sdb: sdb1

Oct 8 00:12:13 omicron kernel: sda: sda1

Thinking that means it have found a partition table for the drives.

 

In the /var/log directory I have the following files:

smart_mid_after_zero1.sda

smart_mid_after_zero1.sdb

smart_mid_preread1_sda

smart_mid_preread1_sdb

smart_start_sda

smart_start_sdb

 

in the /tmp directory I have the following files:

disk_not_normal - empty

preclear_assigned_disks1 - empty

preclear_stat_sda - sda|N post_read_err |Post-Read. 66% @ 49MB/s (28:30:55)

preclear_stat_sdb - sdb|N post_read_err |Post-Read. 70% @ 50MB/s (28:30:19)

read_speedsda

read_speedsdb

smart_mid_after_zero1_sda

smart_mid_after_zero1_sdb

smart_mid_pending_reallocate_sda

smart_mid_pending_reallocate_sdb

smart_mid_preread1_sda

smart_mid_preread1_sdb

smart_start_sda

smart_start_sdb

 

 

When I check the preclear script whether the disks are precleared or not I get the following:

 

/dev/sda1

Partition 1 does not end on cylinder boundry

Partition 1 does not start on physical sector boundry

 

DISK /dev/sda IS PRECLEARED

 

The same goes for the sdb drive.

 

So my question.

I still need to run the preclear 2 more times (recommended 3 times for new drives as per beginners guide)

 

Do I just start over again or is there anything I've missed or what should I do?

You probably ran out of free memory. (The post-read processes did not complete.)

 

I suggest using the -r -w and -b options of the preclear script to limit the memory used.

 

See

http://lime-technology.com/forum/index.php?topic=21291.msg192198;topicseen#msg192198

and

http://lime-technology.com/forum/index.php?topic=22404.msg198618#msg198618

 

I'd run another cycle on each.

preclear_disk.sh -A -b 500 -w 32768 -r 32768 -M 4 /dev/sda

preclear_disk.sh -A -b 500 -w 32768 -r 32768 -M 4 /dev/sdb

 

The preclear-signature is very similar to a partition  entry in the master-boot-record, do not worry about the messages you saw from some utilities or the partition device entries.  They are expected.

 

 

That's what I though, though I have 4GB of memory thought it would be enough.

I read the second thread you linked but I couldn't really figure out how to use the -b-w commands.

 

I am going to run the following now:

 

./preclear_disk -b 500 -r 32768 -w 32768 -M 4 -c 2 /dev/sda

./preclear_disk -b 500 -r 32768 -w 32768 -M 4 -c 2 /dev/sdb

 

since I am guessing you meant -r and -w not -w, guess it's just a typo : )

Also I am skipping the -A since I already have that configured in the GUI.

And I read somewhere that ALL drives over 2.2TB MUST have a protected GPT which means the -A or -a is irrelevant as it won't change it anyway.

 

Cheers for the help, I'll let it run for like forever and get back with input and hopefully mark it as solved

That's what I though, though I have 4GB of memory thought it would be enough.

I read the second thread you linked but I couldn't really figure out how to use the -b-w commands.

 

I am going to run the following now:

 

./preclear_disk -b 500 -r 32768 -w 32768 -M 4 -c 2 /dev/sda

./preclear_disk -b 500 -r 32768 -w 32768 -M 4 -c 2 /dev/sdb

 

since I am guessing you meant -r and -w not -w, guess it's just a typo : )

Also I am skipping the -A since I already have that configured in the GUI.

And I read somewhere that ALL drives over 2.2TB MUST have a protected GPT which means the -A or -a is irrelevant as it won't change it anyway.

 

Cheers for the help, I'll let it run for like forever and get back with input and hopefully mark it as solved

Correct on all counts...    Also correct, -a or -A is ignored on drives over 2.2TB.

 

Joe L.

P.S. It was a typo... sorry, but you figured it out.

I've run 2 preclear cycles now, one of the drives are done and the other one have 2% left. So should not be any issues there.

 

Thanks a bunch for the help.

Marking this as solved.