January 22, 200818 yr So I have the strangest problem I've ever seen so far. Here's the background: Disk2 starting acting up. I had a red ball next to it and it would sometimes crash and hang when I tried to copy files to it. I was able to copy everything over to disk3 (which happened to be empty). With the help of J.L., I deleted disk2 and rebuilt the array and started a parity-sync. Everything seemed to be fine. But.... My Network Places still shows the old array disks and shares. If I manually navigate to Tower, then it shows the correct disks and shares, but disk 2 is still the OLD disk 2. The new disk 2 should have completely different data on it. Both disk 2 and disk 3 are showing the same data. I know for a fact that I removed the correct disk. I double checked the serial numbers. Also, in the web interface, the disk space usage is showing an obvious discrepancy. Even though windows is showing the exact same data on both disk2 and disk3, unraid is showing a difference of over 120 gigs! Confusing and bizarre. What is going on? Parity-sync is at 88% Will it be different when the parity-sync is done? I can't imagine that it would be. The disks are mounted and they should be showing the correct data. I'm at a loss.
January 22, 200818 yr Author I plugged the old disk2 back in just to make sure it was the correct disk that I pulled. It was. I think somehow the parity rebuilt the old array even though I stopped it and used Restore. I'm at a loss as to what to do now. Maybe my data is gone, but I don't know. I'm a linux noob, so I can't figure out how to view the drives from within telnet. I tried using the "cd" command but I obviously don't know how to use it. The fact that there is such a huge discrepancy in the storage usage, makes me think that the drive data is still there but just not being read properly for some reason. Any help would be great. Thanks.
January 22, 200818 yr Author Ok...so I figured out how to use the "cd" command in the telnet window. My files are there!!! So, what in the hell is going on? Why is windows still reading the old array? I've rebooted all of my machines several times. This makes no sense.
January 22, 200818 yr Author Just to add one more quirk into the mix. I can pull up the "movies" share and my hidden files are there. They play fine even though windows thinks they don't exist and that completely different files are there. Bizzaro.
January 22, 200818 yr Maybe it is just me, but Im not really sure as what you did. deleted disk 2, is that deleting the content or removing disk 2 from the array. Did you put in a replacement or rebuild parity on old disk 2? If disk 2 is screwed up and cant write to it, why would you expect that you can delete stuff from it.. When you say disk 2 is still old disk 2 are you then talking about a physical disk, a logical disk or that the content on a new physical disk 2 with the content of the old disk 2, before you thought you deleted it. All in all Im not totally clear as of the symptoms you describe and what you atually did. /Rene
January 22, 200818 yr Author Understandable. It is really confusing. So disk2 was acting strange. I had a red ball next to it. Random crashes. etc. I was able to copy the files off of disk2 onto disk3 (disk3 was empty). I didn't want to replace disk2 with another disk. The array was big enough. I just wanted to get rid of it. Instead of having 11 disks + parity, I now have 10 disks + parity. So here are the steps I did: 1. Power down the server 2. Physically remove disk 2 3. Power up the server. 4. Assign the disks - disk 11 was moved to disk 1 - disk 1 was moved to disk 2 5. Restored the array 6. Began parity-sync Symptoms I have now: 1. Disks 1 through 11 still show up in My Network Places (even though obviously disk 11 is not accessible) 2. Disk 2 is still showing the old data. It is readable and playable. 3. Disk 2 should show the data that was on my old disk 1 4. Navigating to the tower in windows explorer shows disks 1 through 10 (disk 11 isn't there) but it still shows the old disk 2 data 5. However, my shares are correct. The "missing" data is there in the share and plays just fine. 6. Within telnet, the files are how they should be. Everything is in it's place. The old disk 1 data is there on the new disk 2. So, Unraid is apparently working fine, but windows is reading everything completely screwed up. I've rebooted everything. I tried to re-establish a network using the "setup a home network" windows wizard. The only thing I haven't done is reboot my router and switch. That will probably be next.
January 22, 200818 yr Author One thing I forgot to mention: XBMC is functioning the exact same way as Windows. I can access the files through my share, but the actual disks show up as the old array (minus disk 11). So it isn't just a Windows problem.
January 25, 200818 yr Author Hmmm.... So nobody has any idea? I've tried taking the array down and restoring AGAIN, but I got the same results. I'm completely at a loss as to what to do.
January 25, 200818 yr When you are "restoring," are you stopping the array... then checking the checkbox stating "I'm Sure," then pressing the button marked "Restore" which says it will initialize the array and mark all drives as new, but not delete data... or Are you starting the array and having it restore the contents of the missing drive? and then pressing the "Check" button to check parity. In neither case do you need to do an additional "Parity Sync" If using the "Restore" button it will use the existing data drives and forget about any old drives you used to have connected. It will do a full parity calculation and write that to the parity drive. Looking at your server right now... Does your parity drive have a LOT of writes to it compared to reads, or a lot of reads compared to writes. If the first, you probably did use the "restore" button. If the latter, it used existing paritiy to rebuild your old disk 2.
January 25, 200818 yr Author The first option. I'm doing restore and checking "i'm sure" etc etc Parity drive has 62 reads and 69 writes. I am just using the terminology that the web interface uses. It calls it a parity-sync when I do the restore.
January 25, 200818 yr The first option. I'm doing restore and checking "i'm sure" etc etc Parity drive has 62 reads and 69 writes. I am just using the terminology that the web interface uses. It calls it a parity-sync when I do the restore. You must have re-started your array recently, since the number of reads or writes to your parity drive seem to be just those from the last files you moved onto your server. I know the terminology is confusing... I wish the "Restore" button was labeled "Initialize Array" since that is what it does (and what the description says it does). At least one person pressed it when replacing a failed drive expecting it to restore their data... it didn't, and instead erased their parity drive's memory of the failed drive ever having been in their array. Now what they wanted at all. At this time, are all your drives on-line? (all indicators "green") If nothing else, attach a syslog. Perhaps it will give somebody a clue. Joe L.
January 25, 200818 yr Author Everything is green. Inside Unraid, everything is as it should be. But viewing from windows or XBMC, everything is completely wonky. Edit: I restored one more time today just for giggles. You'll notice that in the syslog. Didn't change a thing though of course.
January 26, 200818 yr Sounds to me like the problem is in your Windows machines. You mentioned Windows Network Places, which I don't believe automatically refreshes, so you may see some improvement after rebooting each of your other devices, and letting them rebuild their 'Network Places'.
January 26, 200818 yr Sounds to me like the problem is in your Windows machines. You mentioned Windows Network Places, which I don't believe automatically refreshes, so you may see some improvement after rebooting each of your other devices, and letting them rebuild their 'Network Places'. A bizarre related item. I have had similar issues with network places in XP. Shares I created months ago still show up, even though they are long gone. Yesterday, everything reset itself in the midst of working in Windows. No, I didn't reboot, so I have no idea what happened. I previously tried refreshing and a bunch of other things - no change. Bill
January 26, 200818 yr Author Like I said above, it isn't just windows. The same problem exists in XBMC.
January 26, 200818 yr Author Also, I've rebooted everything at least 10 times now, including my routers and switches. I've rebuilt the array 4 times. The problem has persisted for days. So, I don't think it's a matter of things of just needing to be refreshed.
January 26, 200818 yr Try this: - open My Network Places - in the Network Tasks panel click on 'View workgroup computers' - in the list that shows up, click your server You should now see a list of your shares - does everything look right?
January 26, 200818 yr ... I know the terminology is confusing... I wish the "Restore" button was labeled "Initialize Array" since that is what it does (and what the description says it does). At least one person pressed it when replacing a failed drive expecting it to restore their data... it didn't, and instead erased their parity drive's memory of the failed drive ever having been in their array. Now what they wanted at all. ... Joe L. The idea behind the 'Restore' term was that it would "Restore" the configuration to it's initial default state (that of all drives appearing "new"), ie, "restoring" the Flash configuration to a null state. Didn't want to use "Initialize" because that has other connotations - someone might think their drives are being initialized. But you raise a good point that some might think they are restoring their data. I think originally it was "Reset array config" - perhaps we'll go back to that terminology.
January 26, 200818 yr Author Tom, When I navigate to the Tower via "view workgroup computers" this is what I see: 1. 10 drives are there (not 11 like in network places) 2. But these 10 drives are not correct. Disk 2 is still displaying the old data, not the new. 3. User shares are correct.
January 26, 200818 yr "network places" is by default showing "once was, maybe it still is, information". It is a collection of networkshares that Windows has seen previously, and keeps in a handy place so you can access them fast. To clear it out, just select the shares and delete them. If you are afraid that you will delete the complete content of your server, shut down the server first. /Rene
January 27, 200818 yr ... I know the terminology is confusing... I wish the "Restore" button was labeled "Initialize Array" since that is what it does (and what the description says it does). At least one person pressed it when replacing a failed drive expecting it to restore their data... it didn't, and instead erased their parity drive's memory of the failed drive ever having been in their array. Now what they wanted at all. ... Joe L. The idea behind the 'Restore' term was that it would "Restore" the configuration to it's initial default state (that of all drives appearing "new"), ie, "restoring" the Flash configuration to a null state. Didn't want to use "Initialize" because that has other connotations - someone might think their drives are being initialized. But you raise a good point that some might think they are restoring their data. I think originally it was "Reset array config" - perhaps we'll go back to that terminology. Perhaps then... "Set a New Drive Configuration" This will configure the array and compute parity based on the drives currently assigned on the drive assignment page. I like "Set" instead of "Reset" since it is not resetting anything... it is Setting the config based on the currently assigned drives. Joe L.
January 27, 200818 yr Author "network places" is by default showing "once was, maybe it still is, information". It is a collection of networkshares that Windows has seen previously, and keeps in a handy place so you can access them fast. To clear it out, just select the shares and delete them. If you are afraid that you will delete the complete content of your server, shut down the server first. /Rene Ok...but I just want to reiterate that XBMC is having the same problem. So it isn't just windows.
January 27, 200818 yr Please type this command in a telnet session and capture & post it's output: cat /etc/samba/smb.shares
January 27, 200818 yr Author Please type this command in a telnet session and capture & post it's output: cat /etc/samba/smb.shares Tower login: root [Disconnect bypassed -- root login allowed.] Linux 2.6.22.5. root@Tower:~# cat /etc/samba/smb.shares #everything wide open security = SHARE guest account = root guest ok = Yes guest only = Yes path = /boot read only = No force user = root map archive = Yes map system = Yes map hidden = Yes create mask = 0711 directory mask = 0711 [disk1] path = /mnt/disk1 read only = No force user = root map archive = Yes map system = Yes map hidden = Yes create mask = 0711 directory mask = 0711 [disk2] path = /mnt/disk2 read only = No force user = root map archive = Yes map system = Yes map hidden = Yes create mask = 0711 directory mask = 0711 [disk3] path = /mnt/disk3 read only = No force user = root map archive = Yes map system = Yes map hidden = Yes create mask = 0711 directory mask = 0711 [disk4] path = /mnt/disk4 read only = No force user = root map archive = Yes map system = Yes map hidden = Yes create mask = 0711 directory mask = 0711 [disk5] path = /mnt/disk5 read only = No force user = root map archive = Yes map system = Yes map hidden = Yes create mask = 0711 directory mask = 0711 [disk6] path = /mnt/disk6 read only = No force user = root map archive = Yes map system = Yes map hidden = Yes create mask = 0711 directory mask = 0711 [disk7] path = /mnt/disk7 read only = No force user = root map archive = Yes map system = Yes map hidden = Yes create mask = 0711 directory mask = 0711 [disk8] path = /mnt/disk8 read only = No force user = root map archive = Yes map system = Yes map hidden = Yes create mask = 0711 directory mask = 0711 [disk9] path = /mnt/disk9 read only = No force user = root map archive = Yes map system = Yes map hidden = Yes create mask = 0711 directory mask = 0711 [disk10] path = /mnt/disk10 read only = No force user = root map archive = Yes map system = Yes map hidden = Yes create mask = 0711 directory mask = 0711 [Data] path = /mnt/user/Data read only = No force user = root map archive = Yes map system = Yes map hidden = Yes create mask = 0711 directory mask = 0711 [Movies] path = /mnt/user/Movies read only = No force user = root map archive = Yes map system = Yes map hidden = Yes create mask = 0711 directory mask = 0711 [Music] path = /mnt/user/Music read only = No force user = root map archive = Yes map system = Yes map hidden = Yes create mask = 0711 directory mask = 0711 [TV Shows] path = /mnt/user/TV Shows read only = No force user = root map archive = Yes map system = Yes map hidden = Yes create mask = 0711 directory mask = 0711 [TV shows] path = /mnt/user/TV shows read only = No force user = root map archive = Yes map system = Yes map hidden = Yes create mask = 0711 directory mask = 0711 [disk2] path = /mnt/user/disk2 read only = No force user = root map archive = Yes map system = Yes map hidden = Yes create mask = 0711 directory mask = 0711 root@Tower:~#
January 27, 200818 yr Author "network places" is by default showing "once was, maybe it still is, information". It is a collection of networkshares that Windows has seen previously, and keeps in a handy place so you can access them fast. To clear it out, just select the shares and delete them. If you are afraid that you will delete the complete content of your server, shut down the server first. /Rene I deleted all of the shares in "My Network Places" with the server turned off. When I rebooted, all of the shares are gone. I can still get to the Tower via "View Workgroup Computers"
Archived
This topic is now archived and is closed to further replies.