James P Posted November 25, 2010 Share Posted November 25, 2010 Last week I built an Unraid following the recipe for the 20 drive beast. I read that preclearing would save a lot of time, but it was near the end of the week, I didn't have an account to download the script, and the tutorial said you didn't need to, so I skipped the preclearing step. I initially put in four 2TB drives and set it up with one as the parity drive and the other 3 as data drives. The initial formatting/clearing of these drives went quite quickly so I put in 8 more drives (four 1TB drives and four 500GB drives) and left it to format them over the weekend. It has now been almost a week and it says it is still formatting them. I'm not in a hurry, but I am starting to wonder if it is really still formatting/clearing the disk and if so, how long it will take. Am I missing something simple? (I hit the refresh button on the web interface a lot, but not much seems to change). The noise is driving me crazy and I would like to turn it off and move it to another room, but not if it is almost done, is there a way to tell? I don't have any data on it yet so I am open to starting over if that will save me some time. Link to comment
aiden Posted November 26, 2010 Share Posted November 26, 2010 Well, you said you're not in a hurry. If you don't have any data saved in your array then it's a no-brainer to me. You can preclear all your drives simultaneously and be done with 2 full cycles in 5 days for all your drives. At that point you will be able to identify which drives, if any, are having issues. Link to comment
kenoka Posted November 26, 2010 Share Posted November 26, 2010 If it's clearing the drives, it should have a % done, IIRC. Link to comment
Joe L. Posted November 26, 2010 Share Posted November 26, 2010 Formatting only takes a few minutes once a disk is cleared. Clearing a disk (by unRAID) usually runs at near the full sped a disk can be written to... but I've seen deadlocks when multiple disks are being cleared at the same time. A clear of a single disk by unRAID should take less than 12 hours. (2TB = 2000 GB = 2000000 MB. At 50 MB/s that would be just over 11 hours) The way to know what is happening is to look in the syslog. Instructions on how to capture it are in the wiki under "troubleshooting" Joe L. Link to comment
James P Posted November 26, 2010 Author Share Posted November 26, 2010 Where would the % done be? Link to comment
Joe L. Posted November 26, 2010 Share Posted November 26, 2010 Where would the % done be? In the syslog... like I said. Joe L. Link to comment
James P Posted November 26, 2010 Author Share Posted November 26, 2010 Sorry Joe. You're not the one who suggested there should be a % done. That sounded easier than looking in the syslog. I have looked at the syslog now. The last entries say that the drives were spun down a week ago and so it looks like it hasn't been doing anything for the last week. Link to comment
Joe L. Posted November 26, 2010 Share Posted November 26, 2010 Sorry Joe. You're not the one who suggested there should be a % done. That sounded easier than looking in the syslog. I have looked at the syslog now. The last entries say that the drives were spun down a week ago and so it looks like it hasn't been doing anything for the last week. Did it show the series of "Clearing ..." messages? Joe L. Link to comment
James P Posted November 26, 2010 Author Share Posted November 26, 2010 The syslog is very long (attached) and I couldn't find anything about clearing in it. I've turned it off for now and will try to start over later today. I am a little discouraged that it went for a week not doing anything, but not giving me much indication that there was a problem. James syslog.zip Link to comment
Joe L. Posted November 26, 2010 Share Posted November 26, 2010 I can see the initial device configuration here: Nov 18 16:05:48 rover kernel: md: import disk0: [8,144] (sdj) WDC WD20EARS-00M WD-WCAZA1168874 offset: 63 size: 1953514552 Nov 18 16:05:48 rover kernel: md: import disk1: [8,160] (sdk) WDC WD20EARS-00M WD-WMAZA1029822 offset: 63 size: 1953514552 Nov 18 16:05:48 rover kernel: md: import disk2: [8,176] (sdl) WDC WD20EARS-00M WD-WMAZA1232815 offset: 63 size: 1953514552 Nov 18 16:05:48 rover kernel: md: import disk3: [8,192] (sdm) WDC WD20EARS-00M WD-WMAZA1313766 offset: 63 size: 1953514552 I then see a parity check beginning. Apparently you did not have a valid parity disk at this time. (or you did not stop the array before shutting it down the previous time) The array is starting a full parity calculation. Nov 18 16:05:48 rover kernel: md: recovery thread woken up ... Nov 18 16:05:48 rover kernel: md: recovery thread syncing parity disk ... I see the ethernet link being established at 100MB/s. (I'm assuming you do not have a Gigabyte router/switch) Nov 18 16:05:50 rover kernel: e1000e: eth0 NIC Link is Up 100 Mbps Full Duplex, Flow Control: RX/TX I see you assigning the new disk 4: Nov 18 16:07:54 rover kernel: md: unRAID driver 0.95.4 installed Nov 18 16:07:54 rover kernel: md: import disk0: [8,144] (sdj) WDC WD20EARS-00M WD-WCAZA1168874 offset: 63 size: 1953514552 Nov 18 16:07:54 rover kernel: md: import disk1: [8,160] (sdk) WDC WD20EARS-00M WD-WMAZA1029822 offset: 63 size: 1953514552 Nov 18 16:07:54 rover kernel: md: import disk2: [8,176] (sdl) WDC WD20EARS-00M WD-WMAZA1232815 offset: 63 size: 1953514552 Nov 18 16:07:54 rover kernel: md: import disk3: [8,192] (sdm) WDC WD20EARS-00M WD-WMAZA1313766 offset: 63 size: 1953514552 Nov 18 16:07:54 rover kernel: md: import disk4: [8,80] (sdf) WDC WD10EADS-00L WD-WCAU46091239 offset: 63 size: 976762552 Nov 18 16:07:54 rover kernel: md: disk4 new disk and here for disk5 Nov 18 16:08:00 rover kernel: md: unRAID driver 0.95.4 installed Nov 18 16:08:00 rover kernel: md: import disk0: [8,144] (sdj) WDC WD20EARS-00M WD-WCAZA1168874 offset: 63 size: 1953514552 Nov 18 16:08:00 rover kernel: md: import disk1: [8,160] (sdk) WDC WD20EARS-00M WD-WMAZA1029822 offset: 63 size: 1953514552 Nov 18 16:08:00 rover kernel: md: import disk2: [8,176] (sdl) WDC WD20EARS-00M WD-WMAZA1232815 offset: 63 size: 1953514552 Nov 18 16:08:00 rover kernel: md: import disk3: [8,192] (sdm) WDC WD20EARS-00M WD-WMAZA1313766 offset: 63 size: 1953514552 Nov 18 16:08:00 rover kernel: md: import disk4: [8,80] (sdf) WDC WD10EADS-00L WD-WCAU46091239 offset: 63 size: 976762552 Nov 18 16:08:00 rover kernel: md: disk4 new disk Nov 18 16:08:00 rover kernel: md: import disk5: [8,96] (sdg) WDC WD10EADS-00L WD-WCAU47825094 offset: 63 size: 976762552 Nov 18 16:08:00 rover kernel: md: disk5 new disk and here for disk6 Nov 18 16:08:10 rover kernel: md: import disk0: [8,144] (sdj) WDC WD20EARS-00M WD-WCAZA1168874 offset: 63 size: 1953514552 Nov 18 16:08:10 rover kernel: md: import disk1: [8,160] (sdk) WDC WD20EARS-00M WD-WMAZA1029822 offset: 63 size: 1953514552 Nov 18 16:08:10 rover kernel: md: import disk2: [8,176] (sdl) WDC WD20EARS-00M WD-WMAZA1232815 offset: 63 size: 1953514552 Nov 18 16:08:10 rover kernel: md: import disk3: [8,192] (sdm) WDC WD20EARS-00M WD-WMAZA1313766 offset: 63 size: 1953514552 Nov 18 16:08:10 rover kernel: md: import disk4: [8,80] (sdf) WDC WD10EADS-00L WD-WCAU46091239 offset: 63 size: 976762552 Nov 18 16:08:10 rover kernel: md: disk4 new disk Nov 18 16:08:10 rover kernel: md: import disk5: [8,96] (sdg) WDC WD10EADS-00L WD-WCAU47825094 offset: 63 size: 976762552 Nov 18 16:08:10 rover kernel: md: disk5 new disk Nov 18 16:08:10 rover kernel: md: import disk6: [8,112] (sdh) WDC WD10EADS-00L WD-WCAU47890465 offset: 63 size: 976762552 Nov 18 16:08:10 rover kernel: md: disk6 new disk You continued until you assigned all the data disks: Nov 18 16:08:39 rover kernel: md: unRAID driver 0.95.4 installed Nov 18 16:08:39 rover kernel: md: import disk0: [8,144] (sdj) WDC WD20EARS-00M WD-WCAZA1168874 offset: 63 size: 1953514552 Nov 18 16:08:39 rover kernel: md: import disk1: [8,160] (sdk) WDC WD20EARS-00M WD-WMAZA1029822 offset: 63 size: 1953514552 Nov 18 16:08:39 rover kernel: md: import disk2: [8,176] (sdl) WDC WD20EARS-00M WD-WMAZA1232815 offset: 63 size: 1953514552 Nov 18 16:08:40 rover kernel: md: import disk3: [8,192] (sdm) WDC WD20EARS-00M WD-WMAZA1313766 offset: 63 size: 1953514552 Nov 18 16:08:40 rover kernel: md: import disk4: [8,80] (sdf) WDC WD10EADS-00L WD-WCAU46091239 offset: 63 size: 976762552 Nov 18 16:08:40 rover kernel: md: disk4 new disk Nov 18 16:08:40 rover kernel: md: import disk5: [8,96] (sdg) WDC WD10EADS-00L WD-WCAU47825094 offset: 63 size: 976762552 Nov 18 16:08:40 rover kernel: md: disk5 new disk Nov 18 16:08:40 rover kernel: md: import disk6: [8,112] (sdh) WDC WD10EADS-00L WD-WCAU47890465 offset: 63 size: 976762552 Nov 18 16:08:40 rover kernel: md: disk6 new disk Nov 18 16:08:40 rover kernel: md: import disk7: [8,128] (sdi) WDC WD10EADS-00L WD-WCAU46053595 offset: 63 size: 976762552 Nov 18 16:08:40 rover kernel: md: disk7 new disk Nov 18 16:08:40 rover kernel: md: import disk8: [8,16] (sdb) ST3500320AS 5QM28L5P offset: 63 size: 488386552 Nov 18 16:08:40 rover kernel: md: disk8 new disk Nov 18 16:08:40 rover kernel: md: import disk9: [8,32] (sdc) ST3500320AS 9QM7HS55 offset: 63 size: 488386552 Nov 18 16:08:40 rover kernel: md: disk9 new disk Nov 18 16:08:40 rover kernel: md: import disk10: [8,48] (sdd) ST3500320AS 5QM29171 offset: 63 size: 488386552 Nov 18 16:08:40 rover kernel: md: disk10 new disk Nov 18 16:08:40 rover kernel: md: import disk11: [8,64] (sde) ST3500320AS 9QM7HLAN offset: 63 size: 488386552 Nov 18 16:08:40 rover kernel: md: disk11 new disk Then you pressed "Start". Notice this is an un-protected expansion since you do not have valid parity. Nov 18 16:09:10 rover kernel: mdcmd (21): start UNPROTECTED_EXPANSION Nov 18 16:09:10 rover kernel: unraid: allocating 64440K for 1280 stripes (12 disks) Nov 18 16:09:11 rover kernel: md1: running, size: 1953514552 blocks Nov 18 16:09:11 rover kernel: md2: running, size: 1953514552 blocks Nov 18 16:09:11 rover kernel: md3: running, size: 1953514552 blocks Nov 18 16:09:11 rover kernel: md4: running, size: 976762552 blocks Nov 18 16:09:11 rover kernel: md5: running, size: 976762552 blocks Nov 18 16:09:11 rover kernel: md6: running, size: 976762552 blocks Nov 18 16:09:11 rover kernel: md7: running, size: 976762552 blocks Nov 18 16:09:11 rover kernel: md8: running, size: 488386552 blocks Nov 18 16:09:11 rover kernel: md9: running, size: 488386552 blocks Nov 18 16:09:11 rover kernel: md10: running, size: 488386552 blocks Nov 18 16:09:11 rover kernel: md11: running, size: 488386552 blocks You apparently saw the un-formatted drives and pressed the "Format" button to format all the new drives. Since you do not have valid parity it is not necessary to clear the drive, just to format them. This occurred here: Nov 18 16:09:42 rover emhttp: shcmd (99): mkreiserfs -q /dev/md7 2>&1 | logger Nov 18 16:09:42 rover emhttp: shcmd (99): mkreiserfs -q /dev/md11 2>&1 | logger Nov 18 16:09:42 rover emhttp: shcmd (99): mkreiserfs -q /dev/md9 2>&1 | logger Nov 18 16:09:42 rover emhttp: shcmd (99): mkreiserfs -q /dev/md6 2>&1 | logger Nov 18 16:09:42 rover emhttp: shcmd (99): mkreiserfs -q /dev/md5 2>&1 | logger Nov 18 16:09:42 rover emhttp: shcmd (99): mkreiserfs -q /dev/md4 2>&1 | logger Nov 18 16:09:42 rover emhttp: shcmd (99): mkreiserfs -q /dev/md10 2>&1 | logger Nov 18 16:09:42 rover emhttp: shcmd (99): mkreiserfs -q /dev/md8 2>&1 | logger The parity sync was completed here: Nov 19 02:13:16 rover kernel: md: sync done. time=36251sec rate=53888K/sec Nov 19 02:13:16 rover kernel: md: recovery thread sync completion status: 0 Your drives have been ready and waiting for you to use them since then. The web-browser will not refresh itself. You'll need to press "refresh" to see the updated status. Everything looks quite normal. Since you did not have existing parity protection the clearing never was needed. Joe L. Link to comment
James P Posted November 26, 2010 Author Share Posted November 26, 2010 Thanks Joe I pushed the refresh button frequently and it always said it was formatting and didn't give me the option of stopping the array like I did when I assigned the additional drives. I turned the whole machine off for a while and have just turned it back on. The web interface says disks 4 through 11 are unformatted. Last time I clicked the "I'm sure I want to do this" box and the "Format" button and waited a week. I'm reluctant to do this again. I assume there is a better way. James PS. I've attached another syslog incase that helps and no, I don't have a gigabit router/switch yet. I haven't decided which one to get yet. syslogNov26.zip Link to comment
Joe L. Posted November 27, 2010 Share Posted November 27, 2010 If you want to do it by hand, you can type the format commands one by one yourself. They were highlighted in my prior post Log in as root and then type: mkreiserfs -q /dev/md4 mkreiserfs -q /dev/md5 mkreiserfs -q /dev/md6 mkreiserfs -q /dev/md7 mkreiserfs -q /dev/md8 mkreiserfs -q /dev/md9 mkreiserfs -q /dev/md10 mkreiserfs -q /dev/md11 Each will take a minute or two to run. When you have formatted each in turn, stop the array by pressing "Stop" and re-start it by pressing "Start". That should then see they are all formatted and they should all be available for use. /dev/md4 = disk4, /dev/md5 = disk5, etc... Joe L. Link to comment
James P Posted November 27, 2010 Author Share Posted November 27, 2010 Thanks, have a good weekend. James Link to comment
Recommended Posts
Archived
This topic is now archived and is closed to further replies.