topcat989 Posted June 7, 2013 Share Posted June 7, 2013 So I have a new 5.0rc12 box, and upgraded it's parity drive with the 2TB disk (parity) from my 4.7 box, and set for parity sync. Started at 10am on the 6th, it said parity would take est 600 minutes, which is 10 hours. Now it's 15 hours, 45 min later and I still hear the drive writing. But when I try to access the web gui to check progress, it times out. I wouldn't mind it taking so long if I could just access the webgui and see progress. Neither the normal webgui nor unMENU come up. Link to comment
garycase Posted June 7, 2013 Share Posted June 7, 2013 Unfortunately you can't restart emhttp in v5, so there's nothing you can do except a hard shutdown. Did you do a parity check BEFORE replacing the parity drive? I'd wait until it's been at least 24 hours since you started the process before doing a hard shutdown => hopefully you'll find that the parity sync completed, so you have good parity. If not, the parity check that will be automatically started on your next reboot will fail miserably !! Note that it may simply be taking a lot longer than the initial "guess" due to the wide mix of sizes you have in the array (this causes notably slower parity checks). Link to comment
dgaschk Posted June 7, 2013 Share Posted June 7, 2013 Try to get a syslog before shutting down. Link to comment
topcat989 Posted June 7, 2013 Author Share Posted June 7, 2013 welp looks like I paniced over nothing, or it was something and seems to have resolved itself. This morning I no longer hear the drive whirring and writing, so just for the hell of it I try the webgui- it comes right up, with parity sync in process. Avg speed around 80mb/s. Maybe it cleared the drive and then started the parity sync? Either way, glad I let it run instead of hard reboot. Link to comment
garycase Posted June 7, 2013 Share Posted June 7, 2013 Sounds like all is well -- post a note when it completes Link to comment
JonathanM Posted June 7, 2013 Share Posted June 7, 2013 Sounds like all is well -- post a syslog captured when it completes FTFY It would be nice to see what was happening in the syslog during all that time the server was incommunicado. Link to comment
topcat989 Posted June 7, 2013 Author Share Posted June 7, 2013 Sounds like all is well -- post a note when it completes Parity sync just completed, now doing a parity check.........est time 550 minutes arrg. Wait and wait and more wait. People who have to move multi TB or add multi TB disks must have their servers down for days... Link to comment
garycase Posted June 7, 2013 Share Posted June 7, 2013 Sounds like all is well -- post a note when it completes Parity sync just completed, now doing a parity check.........est time 550 minutes arrg. Wait and wait and more wait. People who have to move multi TB or add multi TB disks must have their servers down for days... Except for the time needed to do a parity sync, and then occasional parity checks, you can eliminate the long wait times by pre-clearing the drives. The pre-clears take a LONG time, but they have no impact on the array. Of course there's nothing you can really do to speed things up when you have to copy TB's of data ... that is simply a waiting game. Link to comment
topcat989 Posted June 7, 2013 Author Share Posted June 7, 2013 well I'm not out of the woods yet. Came back and after parity check this is what I've got: I was planning on replacing the 300GB here with the 1.5TB drive. (I populated the 5.0 array first with the 500GB and 300GB drives in order to have space for the data off a 750GB drive in my old server, I moved the data off it, then moved it to this array, etc) So....what do I do? How to replace the 300GB drive and not lose data on it if parity is not to be trusted? move the data off the 300gb drive onto the free 750gb drive (disk 1)? then remove 300gb drive, initconfig, then parity sync+check? Just replace the 300gb drive now with the 1.5tb drive? Arrrg, I've had unRAID for almost 6 years but I've never really tweaked with it except at the beginning. Got it up and going, and just updated thru 4.7. It just sat there and ran. I wouldn't have even updated to 5.0 except that I knew I would have to populate this smaller server with 2TB> drives soon. Sorry for you guys having to hold my hand and help me across the street, but I sure do appreciate it Link to comment
JonathanM Posted June 8, 2013 Share Posted June 8, 2013 At this point, I suspect the 300GB drive was in the process of failing while you were trying to build parity, which means the current parity probably will not reconstruct the data from the 300 correctly, which means no rebuilding, and no copying from the virtual drive. If you can access the data from the 300GB and it looks good, then by all means back it up now. The fact that the drive is red balled means you aren't reading from the drive anyway, you are reading from the parity and all the other disks to simulate the 300GB drive. Do you have a backup of the data off the 300GB elsewhere? Is the previous parity drive still in the same condition it was before you started the upgrade, and has there been anything written to any drive in the array since the start of the parity upgrade? I suspect you are about to lose the 260GB of data that was on that drive. Link to comment
garycase Posted June 8, 2013 Share Posted June 8, 2013 The system would not have successfully completed the parity sync operation if it had red balled the disk 4 during that operation ... so you PROBABLY have a good parity disk. IF that's the case, you can simply replace the 300GB drive with the 1.5TB, and Start the array. It will immediately commence a rebuild of the disk onto the new drive ... and when it's done you'll have all the data that was on the 300GB disk plus another 1.2TB of storage space IF for some reason the parity check completed without recognizing the defective drive, then it's true that the rebuilt drive may not have the correct data ... but if that's the case, most likely the 300GB drive doesn't either. I'd definitely look at the 300GB drive after you remove it (in your eSATA caddy on Windows); and compare it to the data on the rebuilt disk after it's done. Link to comment
topcat989 Posted June 8, 2013 Author Share Posted June 8, 2013 ok I've been reviewing the data on the 300gb drive and it seems to be good (or if it's the array simulating the disk) i've played several HD movies and no hiccups. I'm just copying the data over to free space on HD#1. After that I confirm succesful copy, I'll delete the data off 300gb drive, unasign it, then physically remove it, initconfig, then add the 1.5TB drive as a free space drive (no data rebuild because it's moved off the 300GB drive Yes? No? Link to comment
garycase Posted June 8, 2013 Share Posted June 8, 2013 ok I've been reviewing the data on the 300gb drive and it seems to be good (or if it's the array simulating the disk) i've played several HD movies and no hiccups. I'm just copying the data over to free space on HD#1. After that I confirm succesful copy, I'll delete the data off 300gb drive, unasign it, then physically remove it, initconfig, then add the 1.5TB drive as a free space drive (no data rebuild because it's moved off the 300GB drive Yes? No? NO ... no need to copy the data. If you're "reading" the data from disk #4 from the array, then that means your parity is fine. All you need to do is let it rebuild the drive ... all of the data will be on the new drive If you're reading the data from the physical drive via an eSATA connection to your Windows machine; then keep the drive until you confirm the rebuild goes okay. But if UnRAID is letting you read the drive; and the data looks good; then it will rebuild the drive just fine. That's what fault-tolerance is all about !! Link to comment
Recommended Posts
Archived
This topic is now archived and is closed to further replies.