August 25, 200916 yr I'm sorry if this has been covered on the wiki. I did a quick scan on the wiki and found nothing. So please answer my question. I'm gonna merge my data from the current NAS. I'm gonna use my backup storage to rebuild my storage on the new unRaid nas. However i'm ... running out of back up space. Luckily i have 1 TB drive in the current NAS thats not in any RAID array (i did a quick expansion so i didnt rebuild the array). Hence i'm wondering if i can just place this drive into the new unRAID without doing any.... data merging? The drive is currently formatted with XFS filesystem. Please i just need a quick answer to know if i should take the current NAS apart right now.
August 25, 200916 yr unRAID will want to format the drive so you will lose the data on it when you put it in the server. The best way is to transfer one drive to the machine and start moving stuff to that drive, keep doing this will all the drives and until the data is transferred and then assign the parity drive and do a parity calculation and parity check to make sure everything is working correctly. The only thing you have to make sure about is that you leave the largest drive to be the parity.
August 25, 200916 yr Author Ouch so it seems like unraid only support/work with reiserFS then... (Any other FS support in the future?) Ok so my plan is to get another extra 1tb disk. Format it as reiserFS and copy the data from the xfs disk. Put the new disk in and thats it? I wont turn on the parity untill i got all the data in (and checking thro unraid) Sound good?
August 25, 200916 yr Ouch so it seems like unraid only support/work with reiserFS then... (Any other FS support in the future?) Ok so my plan is to get another extra 1tb disk. Format it as reiserFS and copy the data from the xfs disk. Put the new disk in and thats it? I wont turn on the parity untill i got all the data in (and checking thro unraid) Sound good? It is not just formatted as ReiserFS... it must be formatted as unRAID would have, with a single partition starting at sector 63 to the end of the disk, leaving the entire first cylinder unused, otherwise it will consider it foreign and consider it "Unformatted" Best is to use the preclear_disk.sh script I wrote to clear AND format the drive... Then you can use mkreiserfs /dev/sdX on it. It will then be formatted exactly as needed. You can move the reiserfs disk to your other Linux box to copy the files to it, then move it back... or copy the files over the LAN... your choice. Tom has said he will be adding support for other file systems at some time in the future, but don't hold your breath waiting... there is a fairly major user-interface change planned for unRAID 5.X and I think it will come first.
August 25, 200916 yr Author As you can tell i'm very new to this. Where can i download your script? (it would be very nice if you can include the link in your sig) Also you have a guide on how to run your script? I will break into the following steps: + Put the new disk to my linux box + run your script ( ./preclear_disk.sh /dev/sdX ) + the disk is now ready to create filesystem : mkreiserfs /dev/sdX1 (i assume the disk is now properly formatted and partitioned from sector 63 and only 1 partition?) + copy data + put the disk into the unRaid box. + data is ready for parity?
August 25, 200916 yr As you can tell i'm very new to this. Where can i download your script? (it would be very nice if you can include the link in your sig) I've written far too many add-ons and scripts to have them in my sig. Instead is a link to the wiki. If you searched there, or in this forum, you would have found it. In any case, the script is attached to this post: http://lime-technology.com/forum/index.php?topic=2817.0 Download to your PC, unzip. and move to the root of your flash drive. Then, log in on the system console, type cd /boot to get to where the flash drive is mounted, and you can then run the script (exactly as you described in fact) Also you have a guide on how to run your script? Yes, a very long thread describes it. type: preclear_disk.sh /dev/sdX I will break into the following steps: + Put the new disk to my linux box + run your script ( ./preclear_disk.sh /dev/sdX ) That is exactly the syntax + the disk is now ready to create filesystem : mkreiserfs /dev/sdX1 (i assume the disk is now properly formatted and partitioned from sector 63 and only 1 partition?) + copy data + put the disk into the unRaid box. + data is ready for parity? You've basically got it. In fact, you even knew it was /dev/sdX1... cool. Joe L.
August 28, 200916 yr Author Preclearing process failed .... I was busying backing up all my data to the backup storage. So today i finally tried out 2 new WD Green 1TB that i just bought. The preclearing process was done on Ubuntu 9.04 Server Edition. First, i installed SmartMonTools package so the script can use the "smartctl" command. Then the preclearing process started on one drive. After ... around 6 hrs or so. I logged back into the console though my KVM and saw an error saying " /dev/sdd can not be precleared" Altho i saw all 10 steps were done ... I then checked the disk with fdisk and as suspected the disk was not partitioned. This leads me to think the disk was failed at the write state somehow. I was furious and in a rush so i re-started the preclear process. I should have taken a screenshot to show the exact error. Can you shred some light on this ?
August 28, 200916 yr Author Oops, the correct msg was: /dev/sdX MBR could not be pre-cleared. Total elapsed time was 8:20 hours. Result was the same. No partitioning was done.
August 28, 200916 yr Oops, the correct msg was: /dev/sdX MBR could not be pre-cleared. To see the actual contents of the first 512 bytes on the disk, you can type this command (substituting your correct device name) dd if=/dev/sdb count=1 | od -x -Ad The output will look similar to this (actual values depend on your specific drive geometry): 1+0 records in 1+0 records out 512 bytes (512 B) copied, 0.000512895 s, 998 kB/s 0000000 0000 0000 0000 0000 0000 0000 0000 0000 * 0000448 0000 0000 0000 003f 0000 7af1 aea8 0000 0000464 0000 0000 0000 0000 0000 0000 0000 0000 * 0000496 0000 0000 0000 0000 0000 0000 0000 aa55 0000512 To learn the drive geometry, type this (again, substituting your drive for "sdb"): fdisk -l /dev/sdb The output will look similar to this: Disk /dev/sdb: 1500.3 GB, 1500301910016 bytes 255 heads, 63 sectors/track, 182401 cylinders Units = cylinders of 16065 * 512 = 8225280 bytes Disk identifier: 0x00000000 Device Boot Start End Blocks Id System /dev/sdb1 1 182402 1465138552+ 0 Empty Partition 1 does not end on cylinder boundary. Post both commands output, and I'll have an idea what happened (or did not happen) Joe L.
August 28, 200916 yr Author ~$ sudo dd if=/dev/sdd count=1 | od -x -Ad 0000000 0000 0000 0000 0000 0000 0000 0000 0000 * 0000496 0000 0000 0000 0000 0000 0000 0000 5c5c 0000512 1+0 records in 1+0 records out 512 bytes (512 B) copied, 0.488328 s, 1.0 kB/s Note that since i'm using ubuntu, i used sudo call these commands. (and when i run the script) sudo fdisk -l /dev/sdd Disk /dev/sdd: 1000.2 GB, 1000204886016 bytes 255 heads, 63 sectors/track, 121601 cylinders Units = cylinders of 16065 * 512 = 8225280 bytes Disk identifier: 0x00000000 Disk /dev/sdd doesn't contain a valid partition table
August 28, 200916 yr I have absolutely no idea what would happen under ubuntu. I know I am reading data from /proc that will only exist on unRAID.. but that is only to verify you are not trying to clear a drive that is already assigned to your array. The differences are probably in the way the version of "dd," "bash," and "awk" on ubuntu work. From your output... As you already know... there is no partition table. It would be on the line for addresses starting at 0000448. (your bytes are all zero) The last two bytes are the MBR signature that should be "aa55" you have "5c5c" It looks like the disk was not written to at all. (perhaps some kind of permissions?) as you said, you need to "su" to do anything in ubuntu (also to prevent you from shooting yourself in the foot) I'd do the preclear on unRAID. You can boot up a flash drive that has the preclear_disk.sh script at its top level directory and unRAID loaded. It can be the free version. Use 4.5beta6, as it has the working smartctl. (The missing library is back into place) You do not need to assign any drives, or even to start the unRAID array, you just need to be using unRAID when using preclear_disk.sh Joe L.
August 28, 200916 yr Author I have absolutely no idea what would happen under ubuntu. I know I am reading data from /proc that will only exist on unRAID.. but that is only to verify you are not trying to clear a drive that is already assigned to your array. The differences are probably in the way the version of "dd," "bash," and "awk" on ubuntu work. From your output... As you already know... there is no partition table. It would be on the line for addresses starting at 0000448. (your bytes are all zero) The last two bytes are the MBR signature that should be "aa55" you have "5c5c" It looks like the disk was not written to at all. (perhaps some kind of permissions?) as you said, you need to "su" to do anything in ubuntu (also to prevent you from shooting yourself in the foot) I'd do the preclear on unRAID. You can boot up a flash drive that has the preclear_disk.sh script at its top level directory and unRAID loaded. It can be the free version. Use 4.5beta6, as it has the working smartctl. (The missing library is back into place) You do not need to assign any drives, or even to start the unRAID array, you just need to be using unRAID when using preclear_disk.sh Joe L. This was exactly on my mind. I was thinking of using Unraid or even my live stresslinux to do the preclearing. Altho i dont have any evidence to blame ubuntu, i still think its due to the permission. Its strange that if i run a script with sudo it should then be passed to all the commands in the script. I dont have time to do it today or even this week (i'm going on a trip). It will have to be this weekend. Oh well.
Archived
This topic is now archived and is closed to further replies.