November 3, 201213 yr Simple enough. The preclear drives script couldn't e simpler, but are you serious? I've been clearing 3TB drives for 60 hours! The 1TB drives finished in 8-10hrs. They are 7200RPM 3TB are 5900RPM and so I imagined it would be slower, but this much slower is insane. Then again, it's on step 2/10 @ 60% with a I/o of like 10mb/s.... The 1TB were moving at 100-150 if I recall. Ideas?
November 3, 201213 yr Author At 3 passes it will be two weeks before I can dump data over. ~ 50tb of drives is just going to take forever it seems.
November 3, 201213 yr Author For those of you with pre-clearing- specific systems built. Are you running Unraid in it as well or just Linux and required depends of pre clear script?
November 3, 201213 yr For those of you with pre-clearing- specific systems built. Are you running Unraid in it as well or just Linux and required depends of pre clear script? 10MB/s is really slow... Unfortunately, it might be an indication of some bottleneck in your disk controller or motherboard, or, a disk errors of some kind. Have you examined your syslog for evidence of disk errors? I would expect 50 to 100MB/s rates on most disks. Do you have the BIOS set for emulated legacy mode(emulating an old IDE drive) That would slow things down a whole lot. Is this a n old PCI bus MB? That would limit throughput. Are you doing multiple disks at the same time? That too might tax an older MB. (or even some newer ones if you did not limit the memory used with the -r, -w, and -b options to the preclear script) Joe L.
November 3, 201213 yr Author Joe L. Didn't use any flags on the script. I believe the motherboard is set to AHCI, but I am using no On board controllers. 1015Ms. Haven't checked syslogd yet. I am doing six or seven drives at once. However many ttys were available.
November 3, 201213 yr Author I was getting 100MB/s before Instarted the other drives I think. I believe.
November 4, 201213 yr I had the same problem and after some trouble shooting by running a memory test and checking my unraid flash drive with check disk on my windows PC I found that I had bad memory and my unraid flash drive had some errors on it. So I replaced my system memory and flash drive and had no problems preclearing 6 drives at 80-100 MB/s.
November 4, 201213 yr I just did 3 x 3T Toshiba aka ex Hitachi drives. The preclear times were 28, 30 and 34hrs. These are 7200rpm drives but still yours should not go that much slower. You've got a problem with your system if your drives are going that slow.
November 4, 201213 yr Author Ran memtest for four days straight with no problems before installing Unraid. Unraid flash is on a Sony drive straight from Tom. Scanned it as well. No issues. I am using Corsair XMS3 Memory. I'll swap it for some value Kingston I have laying around. See what happens. I have more of the same motherboard and CPU as well. Only other likely candidate is my cards. Bleah. More of those sitting on my desk but I hate taking this damn case apart.
November 4, 201213 yr Author Nov 4 04:40:01 Hoard syslogd 1.4.1: restart. Nov 4 04:40:05 Hoard emhttp: shcmd (105934): rmmod md-mod |& logger Nov 4 04:40:05 Hoard emhttp: shcmd (105935): modprobe md-mod super=/boot/config/super.dat slots=24 |& logger Nov 4 04:40:05 Hoard emhttp: shcmd (105936): udevadm settle Nov 4 04:40:05 Hoard emhttp: Device inventory: Nov 4 04:40:05 Hoard emhttp: WDC_WD30EZRX-00MMMB0_WD-WCAWZ2718461 (sdb) 2930266584 Nov 4 04:40:05 Hoard emhttp: WDC_WD30EZRX-00MMMB0_WD-WCAWZ2706097 (sdc) 2930266584 Nov 4 04:40:05 Hoard emhttp: ST31000524AS_6VPE4T0Z (sdd) 976762584 Nov 4 04:40:05 Hoard emhttp: ST31000524AS_5VP894K1 (sde) 976762584 Nov 4 04:40:05 Hoard emhttp: ST31000528AS_9VP75LTA (sdf) 976762584 Nov 4 04:40:05 Hoard emhttp: ST31000528AS_9VP93SM3 (sdg) 976762584 Nov 4 04:40:05 Hoard emhttp: ST31000524AS_9VPET4ZJ (sdh) 976762584 Nov 4 04:40:05 Hoard emhttp: shcmd (105937): /usr/local/sbin/emhttp_event driver_loaded Nov 4 04:40:05 Hoard kernel: md: unRAID driver removed Nov 4 04:40:05 Hoard kernel: md: unRAID driver 2.1.4 installed Nov 4 04:40:05 Hoard kernel: read_file: error 2 opening /boot/config/super.dat Nov 4 04:40:05 Hoard kernel: md: could not read superblock from /boot/config/super.dat Nov 4 04:40:05 Hoard kernel: md: initializing superblock Nov 4 04:40:05 Hoard kernel: mdcmd (1): import 0 0,0 Nov 4 04:40:05 Hoard kernel: mdcmd (2): import 1 0,0 Nov 4 04:40:05 Hoard kernel: mdcmd (3): import 2 0,0 Nov 4 04:40:05 Hoard kernel: mdcmd (4): import 3 0,0 Nov 4 04:40:05 Hoard kernel: mdcmd (5): import 4 0,0 Nov 4 04:40:05 Hoard kernel: mdcmd (6): import 5 0,0 Nov 4 04:40:05 Hoard kernel: mdcmd (7): import 6 0,0 Nov 4 04:40:05 Hoard kernel: mdcmd (: import 7 0,0 Nov 4 04:40:05 Hoard kernel: mdcmd (9): import 8 0,0 Nov 4 04:40:05 Hoard kernel: mdcmd (10): import 9 0,0 Nov 4 04:40:05 Hoard kernel: mdcmd (11): import 10 0,0 Nov 4 04:40:05 Hoard kernel: mdcmd (12): import 11 0,0 Nov 4 04:40:05 Hoard kernel: mdcmd (13): import 12 0,0 Nov 4 04:40:05 Hoard kernel: mdcmd (14): import 13 0,0 Nov 4 04:40:05 Hoard kernel: mdcmd (15): import 14 0,0 Nov 4 04:40:05 Hoard kernel: mdcmd (16): import 15 0,0 Nov 4 04:40:05 Hoard kernel: mdcmd (17): import 16 0,0 Nov 4 04:40:05 Hoard kernel: mdcmd (18): import 17 0,0 Nov 4 04:40:05 Hoard kernel: mdcmd (19): import 18 0,0 Nov 4 04:40:05 Hoard kernel: mdcmd (20): import 19 0,0 Nov 4 04:40:05 Hoard kernel: mdcmd (21): import 20 0,0 Nov 4 04:40:05 Hoard kernel: mdcmd (22): import 21 0,0 Nov 4 04:40:05 Hoard kernel: mdcmd (23): import 22 0,0 Nov 4 04:40:05 Hoard kernel: mdcmd (24): import 23 0,0 Nov 4 04:40:05 Hoard emhttp_event: driver_loaded That is from recently, but is that error on super block an issue?
November 5, 201213 yr No, not if you are just setting up the array, as it has not yet been written to the flash drive.
November 5, 201213 yr Author Everything is pre-cleared except a few 3 & 4 TB drives. Took 90 hours to finish the first 3TB. Going to pull those drives and plug into a diff box to clear and test the speed there. Curious if I should be using a 7200RPM drive as parity or is 5900 enough since I would be lucky to get 30MB/s write speed.
November 7, 201213 yr Curious if I should be using a 7200RPM drive as parity or is 5900 enough since I would be lucky to get 30MB/s write speed. Writing to the array is limited by the slower of the rotational speeds of the two disks involved. (unless concurrently writing to two different data drives, then the faster party drive may help some) Therefore , rotational speed usually does make a difference, but only if the data disks are also 7200 RPM. The eventual sustained "write" speed is directly limited by the rotational speed of the disks since a sector must first be read, then written, and the entire (parity and data) disk(s) rotate once in between. 7200 RPM disks will be written to under unRAID roughly a third faster than 5400 RPM disks, but ONLY if BOTH the data disk and the parity disk are spinning at 7200 RPM.
Archived
This topic is now archived and is closed to further replies.