January 29, 200719 yr Tom, The other day I was having trouble with a drive on my 2nd tower and I ran reiserfsck which found some problems so I ran it again with --fix-fixable so it did, but parity was wrong after that. This is to be expected I guess, but is there a way to run reiserfsck without having to re-import the disks and rebuild parity? Simply running the parity check would overwrite the "corrections" that the check made yes? If not it would be nice to manually say to unRAID "This disk is right. fix parity based on this disk." without having to rebuild parity altogether. Thanks.
February 13, 200719 yr With array Started, from telnet session, type this: samba stop [all your shares will disappear from network] umount /dev/md1 ['md1' corresponds to disk1, 'md2' to disk2, etc.] reiserfsck /dev/md1 [answer 'Yes' to prompt] To resume normal operations: mount /dev/md1 /mnt/disk1 [important to match up the 'md1' with 'disk1', 'md2' with 'disk2', etc.] samba start [all shares should again be visible] The reiserfsck can take a long time on a full file system (several minutes). Also re-mounting the disk can take 15 sec or so.
May 19, 200818 yr I followed the intructions Tom posted on how to do this. How will I know when its done?
May 19, 200818 yr It should only take a few minutes to run. I just kept my telnet window open and a few minutes later it updated me on what had or hadn't been fixed.
May 19, 200818 yr It should only take a few minutes to run. I just kept my telnet window open and a few minutes later it updated me on what had or hadn't been fixed. When reiserfsck prompts you to continue, type your answer exactly like requested as "Yes" (Capital "Y", lower case "es") or it will do nothing. When it is done, you will get your command prompt back. If you answer its prompt with all lower case, it will do nothing and return to the prompt immediately. Joe L.
May 19, 200818 yr Yes, the capital "Y" was my mistake. But I'm still a little confused about something. I had meant to run a check on my Parity drive, but I think I actually ran a check on Data disk1. In the dev directory there are only 4 md listed md1, md2, md3, and md5, but I have 5 drives in my system. This makes sense, because now that I think about it, I know that sata ports start with 0, so I should have an md0, but I'm assuming that the parity drive is called something else that I'm not aware of?
May 19, 200818 yr Yes, the capital "Y" was my mistake. But I'm still a little confused about something. I had meant to run a check on my Parity drive, but I think I actually ran a check on Data disk1. In the dev directory there are only 4 md listed md1, md2, md3, and md5, but I have 5 drives in my system. This makes sense, because now that I think about it, I know that sata ports start with 0, so I should have an md0, but I'm assuming that the parity drive is called something else that I'm not aware of? The parity drive does not have a file system on it. It cannot be checked other than by using the "Check" button on the management web-page. It has only "bits" computed based on the equivalent bit positions on the data drive that do have file systems, any attempt to "fix" it using reiserfsck would only damage it, as it has no file system. You will want to run the reiserfsck check on each of your data drives in turn. You system is fine if the checks of md[1234] are ok. Joe L.
May 19, 200818 yr What we need is a flag so that upon bootup the system runs a reiserfsck on all the disks.. like the .autofsck flag in redhat i.e. if [ -f /.autofsck ]; then or at the very least. have some sort of interface where if the file exists and size is > 0... then it contains a list of the devices that should be fixed upon boot up.
May 19, 200818 yr Is it possible for an error on one of my data drives to show up as an error on my parity drive duing a parity check? But for the data drive in question not to have any problems show up in the web interface?
May 19, 200818 yr If you did the reiserfsck on the /dev/sd? device then parity will be out of sync with the data drive. If you did the rieserfsck on the /dev/md? device, then parity should be in sync, if it is out of sync, there might have been some real corruption which would cause the sync issues. After completion of the parity check, if you do it again and errors still exist, there is a larger problem looming.
May 20, 200818 yr This is something that has always confused me... why after so many versions of unRAID isnt this in the GUI. Thats not a dig but checking file system integrity IMO is a core absolutely essential NAS function.
May 20, 200818 yr This is something that has always confused me... why after so many versions of unRAID isnt this in the GUI. Thats not a dig but checking file system integrity IMO is a core absolutely essential NAS function. Let's not bother asking why, (do you really want that answer?) Instead let's impress upon the importance level of this feature. I'll quote something from that thread " In windows there's a way to schedule a chkdsk on reboot. I think that functionality should be within unRAID too." We can go through different iterations as I've made some suggestions on means of growing towards the final polished goal. let's continue the discussion there as this is a feature request. http://lime-technology.com/forum/index.php?topic=2029.0
Archived
This topic is now archived and is closed to further replies.