April 28, 201016 yr Hi all, I bought 2 WD15EARS (Green) 1.5 TB hard drives. I replaced my parity disk with one and add the other as a new data disk. I have done the preclear script without the jumper on 7-8 pins with no problems. All went fine, no errors. When I first try to copy a large file on it ( > 100 GB) I noticed a lot of stuttering and speed inconsistence. It was like the disks where resetting every 30 seconds for many seconds. For smaller files, like 100 MB all is normal. So I searched here and learned about the 7-8 jumpers trick. So tonight, I Unassigned my data disk (not the parity one), stop the array, power down, put the jumper on 7-8 pins, power on, and launch the preclear script. It is terribly slow, the pre-read speed is about 10 MB/s only the pre-read will take like 60 hours ? I also tried to just let unraid rebuild it but again the speed is dropping from 80 KB/sec to like 10 KB/sec and when it is in a speed drop, the server stop responding for a while when I hit refresh. Is it is because I precleared-it and used it wihthout the jumper ? Am I missing a step after the jumper is set ? A special format, a fdisk ? Do I need to run the WD format utility ? or WD align ? Maybe I nedd to delete the partition 1 that is not aligned correctly but how ? Here is my log whe I tried to rebuild the drive..full of errors... Thank you for helping me Ryno log.zip
April 28, 201016 yr No, there is no step after installing the jumper. It sounds to me like you did everything correctly. However, make double-extra sure that you put the jumper on pins 7-8 and not on pins 1-2. It looks like your drives are running a UDMA/133 speeds. For example: Apr 27 19:44:17 fileserver kernel: sdd: sdd1 Apr 27 19:44:17 fileserver kernel: sd 5:0:0:0: [sdd] Attached SCSI disk Apr 27 19:44:17 fileserver kernel: ata6: SATA link up 3.0 Gbps (SStatus 123 SControl 300) Apr 27 19:44:17 fileserver kernel: ata6.00: ATA-8: WDC WD5000AAKS-00YGA0, 12.01C02, max UDMA/133 Apr 27 19:44:17 fileserver kernel: ata6.00: 976773168 sectors, multi 0: LBA48 NCQ (depth 31/32), AA Apr 27 19:44:17 fileserver kernel: ata6.00: configured for UDMA/133 Apr 27 19:44:17 fileserver kernel: scsi 6:0:0:0: Direct-Access ATA WDC WD5000AAKS-0 12.0 PQ: 0 That is true for all 5 of your SATA drives (sda - sde), so you should be seeing these low speeds on all your drives, not just the new ones. I believe this problem is usually cause by bad cabling. First, I would reseat all your SATA cables. If that doesn't work, then you may want to replace them. There's a lot of other errors in your syslog that I don't understand, so hopefully someone else can comment on them.
April 28, 201016 yr You write that you have precleared without the jumper - did you also format it (or in any other way create a partition) without the jumper ? If so, you do need to delete that partition and start over. I'm no expert - I would run the preclear script again with the option to only post-write but I am sure there is an easier way. BUT - your symptoms sound like something else might be a foot, as they say. I did not look at your syslog (again, I am no expert) but you might want to look for power and power cable issues. Insufficient power to the drive could cause it to reset, this might be audible as the drive spins down slightly (or might not be audible). Check cables and PSU.
April 28, 201016 yr Author Thank you all for the replies, I did some more tests tonight. I Tripple check that my jumper is on 7-8 pins : It is OK and I have change my sata cables and my power cables: I does not help: same results, same errors in the syslog. I noticed in my syslog that my problems where on ata2 so I focused more on this drive (sdb) Here are my tests results: I attach you a zip file containing two syslog files - In logWithout.txt, I booted my unraid server without the hard drive being connected. Results: No errors. - In logWOJmp.txt, I booted my unraid server with the drive connected BUT without the Jumper. Results: No errors So without the drive OR with the drive whithout jumper, I have No errors. The only way I have errors is with the jumper. So I also think like MortenSchmidt that because I precleared-it and used it for a while in my array without the jumper: I am stuck with this partition when I set the jumper on. Can someone help me ? Do I need special format, anything ? Other have done it..why cant I ? Logs.zip
April 29, 201016 yr Thank you all for the replies, I did some more tests tonight. I Tripple check that my jumper is on 7-8 pins : It is OK and I have change my sata cables and my power cables: I does not help: same results, same errors in the syslog. I noticed in my syslog that my problems where on ata2 so I focused more on this drive (sdb) Here are my tests results: I attach you a zip file containing two syslog files - In logWithout.txt, I booted my unraid server without the hard drive being connected. Results: No errors. - In logWOJmp.txt, I booted my unraid server with the drive connected BUT without the Jumper. Results: No errors So without the drive OR with the drive whithout jumper, I have No errors. The only way I have errors is with the jumper. So I also think like MortenSchmidt that because I precleared-it and used it for a while in my array without the jumper: I am stuck with this partition when I set the jumper on. Can someone help me ? Do I need special format, anything ? Other have done it..why cant I ? If you type: dd if=/dev/zero of=/dev/sdb count=3 bs=512 You will completely overwrite the partition table and master boot record and the first 2 blocks on the disk. It will, in effect, erase the partition table of anything the operating might think is a file-system. Do this with the jumper off. Then, when you put the jumper on, and the second sector on the disk becomes the first it will not confuse the operating system into thinking some weird partition is on the disk since it was zeroed too. All unRAID will see is zeros in the master boot record. Just be absolutely certain the disk is /dev/sdb, since you will be basically erasing the partition table. Joe L.
April 29, 201016 yr Author Thank you JoeL but it does nothing... Command was sucessful but again, with he jumper back on all the errors came back on syslog (See the zip file attached). It take about 4-5 minutes to boot with all those errors. After that the disk is still accessible but at unusable speed, more stuttering than without etc. See the syslog thank you Ryno logJumperOn.zip
April 29, 201016 yr Author JoeL, Even with all those errors in syslog, I tried to preclear it to see how it responds... It is currently in the preclear step, speed is constant at about 46 MB/S...kind of slow but it is constant (That different from before) = unRAID server Pre-Clear disk /dev/sdb = cycle 1 of 1 = Disk Pre-Read in progress: 1% complete = ( 16,450,560,000 bytes of 1,500,301,910,016 read ) 46.7 MB/s = Disk Temperature: 26C, Elapsed Time: 0:08:57 after I type the preclear command it take like 1.5 minute to have the "Yes" prompt to continue and after I typed "Yes" it took again 1 or 2 minutes to see something append... And it also told me that I do not have a valid partition table for this disk (Because of your command) before typing Yes. But at that speed, I still have a problem !
April 29, 201016 yr after I type the preclear command it take like 1.5 minute to have the "Yes" prompt to continue and after I typed "Yes" it took again 1 or 2 minutes to see something append... That is a symptom of too little memory. Do you have a lot of things running on your server? You might want to use some of the options to the pre-clear script to limit its use of memory. They were first introduced here: http://lime-technology.com/forum/index.php?topic=2817.msg44211#msg44211 And it also told me that I do not have a valid partition table for this disk (Because of your command) before typing Yes. But at that speed, I still have a problem ! This too could be because of too little memory for everything you have running and resource contention.
April 29, 201016 yr Author Currently, my server is only doing the preclear, nothing else. There is nothing running on it exces a task that send me 1 email each day to tell me smart status etc. If I remove the jumper and preclear the same drive all is normal, speed is ok. But when I use it without the jumper and say I write a 100 gig file, at each 20-30 seconds it is like if I pull the network cable for 10 - 15 seconds. If I write a little file, 100 MB all is fine.
April 29, 201016 yr Maybe this will help, latest IBM article on how to work with WD 4K "EARS" drives: http://www.ibm.com/developerworks/linux/library/l-4kb-sector-disks/index.html?ca=dgr-lnxw974KB-Disksdth-LX
April 30, 201016 yr Another user reported earlier a similar problem, EARS drive that works OK without a jumper, albeit slower, but was giving errors with the jumper. If I remember the verdict was bad/buggy firmware on the drive. A RMA fixed his problem and the new drive worked well with jumper. A search may find that thread. I'd say don't waste more time with that drive and try to RMA it.
April 30, 201016 yr Another user reported earlier a similar problem, EARS drive that works OK without a jumper, albeit slower, but was giving errors with the jumper. If I remember the verdict was bad/buggy firmware on the drive. A RMA fixed his problem and the new drive worked well with jumper. A search may find that thread. I'd say don't waste more time with that drive and try to RMA it. maybe purko is referribg to this thread: http://lime-technology.com/forum/index.php?topic=5536.msg55276#msg55276
April 30, 201016 yr I had similar poblems, with jumper on: http://lime-technology.com/forum/index.php?topic=5773.msg54426#msg54426 Anyway, I ran the entire pre clear script and it took about 3 days! After that the disk worked like new.
Archived
This topic is now archived and is closed to further replies.