March 5, 201610 yr I upgraded my parity and adding 5 more drives to my server. When I replaced the parity I did not see a copy file from the old parity for the new parity drive. So I went ahead and started the parity sync but it shows that it will take over 20 days to complete. My question is if I shut off the server to add the norco hot swap and to use for pre clear on new drives will I lose the parity progress?
March 5, 201610 yr I upgrade my parity and adding 5 more drives to my server. When I replaced the parity I did not see a copy file from the old parity for the new parity drive. So I went ahead and started the parity sync but it shows that it will take over 20 days to complete. My question is if I shut off the server to add the norco hot swap and to use for pre clear on new drives will I lose the parity progress? Yes you will lose the progress. But, at 20 days, there's something else going on (most likely you knocked loose a cable somewhere along the line). Diagnostics would tell the tale
March 5, 201610 yr Are you sure you're doing a parity sync? Suspect you're doing a correcting parity check that is finding many errors, that would explain 20 days.
March 5, 201610 yr Author I upgrade my parity and adding 5 more drives to my server. When I replaced the parity I did not see a copy file from the old parity for the new parity drive. So I went ahead and started the parity sync but it shows that it will take over 20 days to complete. My question is if I shut off the server to add the norco hot swap and to use for pre clear on new drives will I lose the parity progress? Yes you will lose the progress. But, at 20 days, there's something else going on (most likely you knocked loose a cable somewhere along the line). Diagnostics would tell the tale I do have 34TB would that make a difference
March 5, 201610 yr I have 73TB and it takes me roughly 24hrs to do a parity sync, so I think there is definitely something else going on with your system. 2.8MB/sec is very slow, that should get up to the high 80's or higher, how long did you leave it for? Perhaps it will speed up and that figure of 20 days will drop?
March 5, 201610 yr Total size of the array doesn't make a difference for time (depending upon your hardware of course) 2.8M / sec is definitely not normal UNLESS you are running all your drives on PCI adapters. Like johnnie said, either its finding a ton of corrections, or like I said you've knocked a cable loose. I'd post your diagnostics
March 5, 201610 yr Author Total size of the array doesn't make a difference for time (depending upon your hardware of course) 2.8M / sec is definitely not normal UNLESS you are running all your drives on PCI adapters. Like johnnie said, either its finding a ton of corrections, or like I said you've knocked a cable loose. I'd post your diagnostics I am using this SUPERMICRO AOC-SASLP-MV8 PCI-Express x4 Low Profile SAS RAID Controller with with 2 Mini SAS to 4 SATA Data Cable supplying sata to 8 drives. The other 6 drives are running off mobo sata ports.
March 5, 201610 yr Total size of the array doesn't make a difference for time (depending upon your hardware of course) 2.8M / sec is definitely not normal UNLESS you are running all your drives on PCI adapters. Like johnnie said, either its finding a ton of corrections, or like I said you've knocked a cable loose. I'd post your diagnostics I am using this SUPERMICRO AOC-SASLP-MV8 PCI-Express x4 Low Profile SAS RAID Controller with with 2 Mini SAS to 4 SATA Data Cable supplying sata to 8 drives. The other 6 drives are running off mobo sata ports. i have a couple of those controllers on my system. I get parity sync speeds in the 80-120MB?sec speed so you should be seeing something similar.
March 6, 201610 yr Author I had installed a rocket raid 2300 for the new norco ss500. I removed that from the system and now I am getting 30mb/sec. I checked all the wires and not loose connections. When I install the rocket raid do I need to do something in the bios? I am getting 20-30mb/sec but to be very honest I know replies say to post diagnostics but I do not know what that means.
March 6, 201610 yr Author Go to tools and click diagnostics, attach the complete zip. Thanks so much, my speeds are back to 3mb/sec. I am attaching the diagnostics zip file. I apologize for being such a technie noob and greatly appreciate the help everyone offers here. tower-diagnostics-20160306-1852.zip
March 7, 201610 yr You are getting CRC errors from at least these three disks, change sata cable/enclosure. Device Model: ST8000AS0002-1NA17Z Serial Number: Z840EWNX 199 UDMA_CRC_Error_Count 0x003e 200 200 000 Old_age Always - 32 Device Model: Hitachi HDS5C3020ALA632 Serial Number: ML0220F30AVLTD 199 UDMA_CRC_Error_Count 0x000a 200 200 000 Old_age Always - 586 Device Model: Hitachi HDS5C3020ALA632 Serial Number: ML4220F3154RDK 199 UDMA_CRC_Error_Count 0x000a 200 200 000 Old_age Always - 520 Also would be a good idea to to an extended SMART test for this one, and if it fails replace it: Device Model: Hitachi HDS5C3020ALA632 Serial Number: ML0220F31BERTD 197 Current_Pending_Sector 0x0022 100 100 000 Old_age Always - 1
March 7, 201610 yr There are no syslog errors from these, but it’s a very short syslog, so keep an eye on these values, if value increases by 2 or more there’s also a problem with sata cable/enclosure, or although less likely, could be a bad port/controller. Device Model: ST3000DM001-9YN166 Serial Number: S1F0KQ7D 199 UDMA_CRC_Error_Count 0x003e 200 199 000 Old_age Always - 214 Device Model: ST4000DM000-1F2168 Serial Number: S3018SVC 199 UDMA_CRC_Error_Count 0x003e 200 197 000 Old_age Always - 132 Device Model: ST4000DM000-1F2168 Serial Number: W300CWZA 199 UDMA_CRC_Error_Count 0x003e 200 197 000 Old_age Always - 16
March 7, 201610 yr Author You are getting CRC errors from at least these three disks, change sata cable/enclosure. Device Model: ST8000AS0002-1NA17Z Serial Number: Z840EWNX 199 UDMA_CRC_Error_Count 0x003e 200 200 000 Old_age Always - 32 Device Model: Hitachi HDS5C3020ALA632 Serial Number: ML0220F30AVLTD 199 UDMA_CRC_Error_Count 0x000a 200 200 000 Old_age Always - 586 Device Model: Hitachi HDS5C3020ALA632 Serial Number: ML4220F3154RDK 199 UDMA_CRC_Error_Count 0x000a 200 200 000 Old_age Always - 520 All 3 of those disks are inside the same Norco SS-500 and I believe it is the oldest Norco I have.
March 8, 201610 yr Author You are getting CRC errors from at least these three disks, change sata cable/enclosure. Device Model: ST8000AS0002-1NA17Z Also would be a good idea to to an extended SMART test for this one, and if it fails replace it: [code]Device Model: Hitachi HDS5C3020ALA632 Serial Number: ML0220F31BERTD 197 Current_Pending_Sector 0x0022 100 100 000 Old_age Always - 1 Ok I have moved around some sata cables and installed some new ones. I went and purchased a new psu also the Corsair CS750M. My parity syncs were still super slow and I ran an extended smart test on this drive and it passed and also ran a smart test on disk 10 that also passed. Problem is now that drive 9 ML0220F31BERTD is showing unmountable. I installed my original parity drive back into the server until I know what to do with this unmountable drive. I am attaching a new diagnostics report but I had to delete a syslog file because the attachment was too large. Thanks again tower-diagnostics-20160308-0736.zip
March 8, 201610 yr Disk9 is not only unmountable, but also disabled, can you explain all the steps you made when changing back to old parity?
March 8, 201610 yr Author Disk9 is not only unmountable, but also disabled, can you explain all the steps you made when changing back to old parity? I just swapped the paritys around and replaced some sata cables. The new drive that I was wanting to use for parity never got past 1% synced. So I figured to just put the old parity back in server. I booted up and now disk 9 is doing this.
March 8, 201610 yr If you haven’t written/deleted anything from the array since you last used this parity disk you can try this: -take a screenshot of current disk assignments -go to tools and click new config -reassign all disks including old parity -before starting array check “parity is already valid”, then start array -check if disk9 mounts This won’t solve your main problem, the UDMA_CRC errors are almost certainly the reason for the very slow parity sync and disable disk, and you have to fix those before attempting another sync, this is just to see if disk9 filesystem is actually ok.
March 8, 201610 yr Author If you haven’t written/deleted anything from the array since you last used this parity disk you can try this: -take a screenshot of current disk assignments -go to tools and click new config -reassign all disks including old parity -before starting array check “parity is already valid”, then start array -check if disk9 mounts This won’t solve your main problem, the UDMA_CRC errors are almost certainly the reason for the very slow parity sync and disable disk, and you have to fix those before attempting another sync, this is just to see if disk9 filesystem is actually ok. Sweet now drive 9 is mounted. It is attempting another parity sync now with speeds around 30MB/sec, should I let it sync for a little?
March 8, 201610 yr It is attempting another parity sync now with speeds around 30MB/sec, should I let it sync for a little? That means you didn't check the “parity is already valid” box before starting the array. If it's going well you can continue but stop if any errors appear on the main page. Maybe also post a syslog an half hour from now so we can see if there still are UDMA_CRC errors.
March 8, 201610 yr Author Well parity sync speed are back to 3MB/sec anfd I attached an updated syslog to see if anything stands out, Thanks tower-syslog-20160308-1053.zip
March 8, 201610 yr You're still getting UDMA_CRC errors, better to cancel the sync and attach full diagnostics.
March 8, 201610 yr Author You're still getting UDMA_CRC errors, better to cancel the sync and attach full diagnostics. My diagnostic it too large to attach. Inside the logs folder there are 2 syslog files on syslog(133kb) and the other syslog.1(128kb). Should I delete one to make attachment smaller?
Archived
This topic is now archived and is closed to further replies.