May 3, 201511 yr Well I am planning to upgrade my 3tb drives to 8tb very very soon. Which software should I use?
May 3, 201511 yr Well I am planning to upgrade my 3tb drives to 8tb very very soon. Which software should I use? Both drive sizes are supported in v5 and above. On the subject of "is it safe". Do you have complete and current backups of all critical data? If there is a problem that results in data loss, could you recover? I am using v6 on my only server, but I have the backups I need offline.
May 3, 201511 yr If I am on v6 and then upgrade the drives, they will rebuilt using xfs. And then I will not be able to go back to v5. Is xfs better than reiser? Should I upgrade via v5 instead? I do not have a backup of my 39tb server.
May 3, 201511 yr If I am on v6 and then upgrade the drives, they will rebuilt using xfs. And then I will not be able to go back to v5. Is xfs better than reiser? Should I upgrade via v5 instead? I do not have a backup of my 39tb server. You can adjust disk settings to change the default filesystem for v6 (options are reiserfs, xfs, and btrfs). Reiserfs allows for disks to be formatted so that they are backwards compatible with v5. Xfs and btrfs are not. As far as differences between file systems, I'd suggest searching this forum. There have been many discussions on the subject but the long and short is that everyone has different opinions.
May 3, 201511 yr Is xfs better than reiser? It's certainly newer and better supported. Reiser is a very stable file system; but is simply not well supported anymore (as you may be aware, the creator of this is in prison for murdering his wife). Probably the biggest disadvantage that may be an issue in a couple years is that it doesn't support partitions > 16TB ...so when drives larger than that are available it won't be a good choice. A minor disadvantage is that it can result in fairly slow writes when a disk is very full (95% or so). Should I upgrade via v5 instead? Probably not -- v6 is stable enough in its NAS functionality that I wouldn't recommend this. Better to upgrade first; THEN add your new drives and use XFS for those drives. I do not have a backup of my 39tb server. Clearly a risk => hopefully there's nothing on your server that you wouldn't mind losing [if that's not the case, you should reconsider your backup strategy (or lack thereof)]
May 3, 201511 yr UnRAID version 5 does support large drives, so there is no absolute need for you to risk all your data with version 6 BETAs. Stick with v5 and add a drive to that.
May 3, 201511 yr If I am on v6 and then upgrade the drives, they will rebuilt using xfs. And then I will not be able to go back to v5. Is xfs better than reiser? Should I upgrade via v5 instead? I do not have a backup of my 39tb server. Drives will not and can not be rebuilt to a different file system regardless of the default file system setting. In fact, if you change the file system on a disk, it will be formatted.
May 3, 201511 yr UnRAID version 5 does support large drives, so there is no absolute need for you to risk all your data with version 6 BETAs. Stick with v5 and add a drive to that. Version 6 is a newer Linux build with support for hardware that may not be compatible with V5. While in general I agree with the sentiment that running beta when you don't need to is not a smart idea, the lengthy beta cycles combined with an active community give users the tools to decide if a beta is stable enough for their use. The beta cycles here are so long and do seem to plateau with stable releases that never go final, and instead have a heap of new features added on top. I would also say that I treat "final" releases with similar caution as a beta. A release is a release, and is susceptible to bugs. But after 30 or so users have been running it for a few weeks, and a set of known issues emerge that I do not consider a risk for me, do I consider it stable enough for my use. If I was going to add a bunch of drives to my server, I would definitely look at the 6 beta to use XFS vs reiserfs. XFS does not seem to suffer from the timeouts and write penalties of reiserfs on large disks as they get fuller and fuller. If you want to preserve a path back to 5.0, stick with reiserfs. But I don't know why you would want to - if 6.0 b15 suddenly were found to have a big issue, we'd move back to b14b, and if that proved a problem back to b12. 6.0 has been in progress for a very long time and there are a lot of stable betas to go back to.
May 3, 201511 yr I think I will go v6. So start it up, stick in a new drive and then it will automatically rebuild in xfs? And I have annoying timeouts on v5!!
May 3, 201511 yr I think I will go v6. So start it up, stick in a new drive and then it will automatically rebuild in xfs? And I have annoying timeouts on v5!! NO!!! IF you put a new drive in to your array and use parity to rebuild the drive it will put whatever file system was on the OLD drive. The reason for this isthat it foes a bit-by-bit rebuilt on the new drive of the data that was on the old drive. See this post about how to convert a drive: http://lime-technology.com/forum/index.php?topic=37490.msg346739#msg346739
May 3, 201511 yr I think I will go v6. So start it up, stick in a new drive and then it will automatically rebuild in xfs? And I have annoying timeouts on v5!! NO!!! As I said just a couple of posts ago: Drives will not and can not be rebuilt to a different file system regardless of the default file system setting. In fact, if you change the file system on a disk, it will be formatted. Note that this means the data will be gone if you change the file system on a disk with data on it.
May 3, 201511 yr Looks like I will be going to v6 and staying on rfs. Upgrading a drive is still the same method right? Updating from v5 to v6 still the same method?
May 3, 201511 yr Looks like I will be going to v6 and staying on rfs. Upgrading a drive is still the same method right? Updating from v5 to v6 still the same method? By upgrading a drive, I assume you mean replacing it with a larger drive and letting it rebuild. Yes that is the same. The rebuilt drive will have the same file system as the original. However, some people, myself included, have converted some or all of their drives to XFS by installing an empty drive or moving everything off a drive to other drives, then formatting to XFS, then moving everything from another drive to that one, repeat. See this sticky for a lot more discussion on getting your data onto another file system. See this wiki for suggestions on best approach for upgrading.
May 4, 201511 yr Experiencing regular reboots on 6B15, system log is showing this Server kernel: mce: [Hardware Error]: Machine check events logged It seems to be a hard reboot as a parity check is initiated (currently running now). I have just recently updated my BIOS, and made a fresh USB 6B15 drive install, with only putting back my user shares and specific config files in the config folder (followed the guide). I cannot say it is specifically B15 making this happen, however it is happening while running it. Syslog attached. *edit - I did remove intel_pstate=disable, so maybe I am actually overheating (someone else had this issue on an Ubuntu forum), ventilation is not the best currently, will look into. *edit 2 - I'm not getting e-mails, because after my re-install I didn't set it back up, doh!! I was infact overheating, had a fan go bad, now fixed. Also I do believe the removal of intel_pstate=disable was a bad idea and created excessive heat, regardless of the recent thought it was only a reporting issue. syslog.txt
May 4, 201511 yr Experiencing regular reboots on 6B15, system log is showing this Server kernel: mce: [Hardware Error]: Machine check events logged Also, I am not getting any e-mail notifications of this happening. It seems to be a hard reboot as a parity check is initiated (currently running now), and I am not getting any e-mails. Normally after a reboot I get ~6 SMART e-mails (need to fix those ) and if a parity check started, I get that also. I have just recently updated my BIOS, and made a fresh USB 6B15 drive install, with only putting back my user shares and specific config files in the config folder (followed the guide). I cannot say it is specifically B15 making this happen, however it is happening while running it. 2 Dockers, 1 VM, and the Community Repositories Plugin. Syslog attached. Syslog attached, No idea if this would help you or not, but you might want to check this out: http://www.advancedclustering.com/act-kb/what-are-machine-check-exceptions-or-mce/ It doesn't look like mcelog comes with unRaid, but that site does offer some suggestions: There most common reason for MCE events to occur are: Memory errors or Error Correction Code (ECC) problems Inadequate cooling / processor over-heating System bus errors Cache errors in the processor or hardware
May 4, 201511 yr Experiencing regular reboots on 6B15, system log is showing this Server kernel: mce: [Hardware Error]: Machine check events logged Also, I am not getting any e-mail notifications of this happening. It seems to be a hard reboot as a parity check is initiated (currently running now), and I am not getting any e-mails. Normally after a reboot I get ~6 SMART e-mails (need to fix those ) and if a parity check started, I get that also. I have just recently updated my BIOS, and made a fresh USB 6B15 drive install, with only putting back my user shares and specific config files in the config folder (followed the guide). I cannot say it is specifically B15 making this happen, however it is happening while running it. 2 Dockers, 1 VM, and the Community Repositories Plugin. Syslog attached. Syslog attached, No idea if this would help you or not, but you might want to check this out: http://www.advancedclustering.com/act-kb/what-are-machine-check-exceptions-or-mce/ It doesn't look like mcelog comes with unRaid, but that site does offer some suggestions: There most common reason for MCE events to occur are: Memory errors or Error Correction Code (ECC) problems Inadequate cooling / processor over-heating System bus errors Cache errors in the processor or hardware Thanks for the info, it does look like I was overheating in one way or another. It is my thought this was a combination of a fan going bad (1 of 3 drawing from the front) and removal of intel_pstate=disable from my go file on re-install that caused these issues. So..... Not related to this thread, sorry.
May 4, 201511 yr Looks like I will be going to v6 and staying on rfs. Upgrading a drive is still the same method right? Updating from v5 to v6 still the same method? I think you may misunderstand => RFS and XFS can "peacefully coexist" with no problem in v6. A NEW disk will automatically be formatted in the current "default" file system (a system option) -- and v6 defaults this to XFS, so unless you change it that's what it will be. The point made earlier is that REBUILDS -- i.e. a rebuild of a failed or replaced disk -- will NOT change the file system that was on the disk being rebuilt. Note that very full (98-99-100%) disks using Reiser do not have any speed issues with READS ... only with writes. So if these disks are essentially static (i.e. full of media that you don't regularly modify/delete/replace) there's no problem with them being as full as you want [16 of the 18 data disks in my media array are either 99 or 100 % full]. Assuming (based on your comments in other threads) that you're planning to move to 8TB drives; I'd first replace your parity drive with an 8TB drive; then upgrade to v6; and then add your first 8TB data drive ... formatted as XFS. All subsequent drives you add will also be XFS. If you want to change your existing drives to XFS (not necessary, but some like to do this), you can simply copy all of the data from one drive to one of your new XFS drives; then reformat that now-empty drive to XFS; and then simply repeat that general process until you're fully converted.
May 4, 201511 yr I get it now. The plan is to move the data from 13 x 3tb drives to 7 x 8tb drives. Like this: replace disk1 3tb with 8tb, rebuild, copy data from disk13 to new disk1, then get remove disk13. And repeat with disk2 and disk12. Why not replace parity drive in v6?
May 4, 201511 yr Why not replace parity drive in v6? Not sure I understand this comment? If you want to move to 8TB data drives then you HAVE to first upgrade the parity drive to 8TB as no data drive can be larger than the parity drive.
May 4, 201511 yr Gary stated that I replace parity in v5 and then upgrade software to v6. And then upgrade drives. I asked, why not replace parity in v6.
May 4, 201511 yr Gary stated that I replace parity in v5 and then upgrade software to v6. And then upgrade drives. I asked, why not replace parity in v6. The order is not that critical. I suspect the reason was just to ensure that all drives were known to be good before trying to make any changes. It also emphasises the point that it is not necessary to move to v6 to use 8TB drives. There should be some other reason you want to move to v6.
May 4, 201511 yr I suspect the reason was just to ensure that all drives were known to be good before trying to make any changes. Absolutely. The FIRST step before making ANY changes should always be to run a parity check to confirm there are no errors -- and if there are any, get them corrected BEFORE making changes. It's not necessary to update parity while still in v5 ... but I'd do it nevertheless. To be precise, I'd do the following: (a) a parity check to confirm all is okay; (b) replace the parity drive and wait for it to rebuild; © another parity check (with the new parity drive); and THEN you can make other changes [add another drive; update to v6; etc.] ... There should be some other reason you want to move to v6. Exactly. If your sole purpose is to add more storage, there's no real need to upgrade. If you want to start switching your drives to XFS; want to use Dockers; want to build a few VM's; etc. then you need to upgrade. There ARE two major NAS functions that have been missing for a LONG time that are finally included in v6 -- UPS support and notifications -- but these are readily available in v5 via UnMenu, so those alone aren't really a compelling reason to upgrade.
May 4, 201511 yr Guys, since you are discussing something I am imminently contemplating doing...Garycase's reply has me in a conundrum. How does one achieve the following items with the lowest amount of stress? 1. Upgrade to V6. 2. Purchase a License (I am currently a Basic license under 5.0.4 - I just have 2 small Share Disks) 3. Move ALL my volumes from RFS to XFS. 4. Install a NEW 6TB Parity Disk. 5. Consolidate 2 1.5 TB Share disks to a single New 4TB disk (which would be XFS, not RFS) 6. Format and reuse the 1.5 TB Share Disks as XFS after I am certain that my files are all intact and everything is working as expected. Up until Garycase's reply, I thought the correct sequence would be: 1. Build a new USB Key with V6 - copying my V5.0.4 Shares and relevant config data to the new USB key. 2. Run a full Parity check - last one run 30 days ago - no errors. 3. Shut down 5.0.4 - Switch USB Keys to V6 variant. 4. Reboot - System should come up EXACTLY like V5 except only accessing 250GB of each share (disk?) 5. If all is well - Purchase License. 6. Shut down server. 7. Install License file on Config subfolder. 8. Reboot. Now all shares should be 100% available. 9. Insert NEW 6TB Drive - Format - Allocate - should become XFS Parity Drive. (XFS? or RFS?) 10. Rerun Parity Check. XFS Parity and RFS Share Disks. 11. If Successful - Insert new 4TB Disk - Format - Allocate - Should be Share Disk 3 and XFS. 12. Initiate copy process to copy/move Share Disk 1 to Share Disk 3. 13. Initiate copy process to copy/move Share Disk 2 to Share Disk 3. 14. Verify all Shares...Share Disks 1 and 2 should empty if data was MOVED. Verify everything works as expected and stable before proceeding. 15. Format Old Parity 2TB Disk for additional storage as XFS. 16. Format Old Share Disks 1 & 2 or reallocate them for other purposes. 17. Purchase and install 240 or 480 GB SSD - setup as XFS or BTRFS for use as Docker Application/Mover Disk 18. Proceed with Dockerization of existing apps (mostly just want to run Universal Media Server), I am still thinking about upgrading Motherboard/CPU/Memory) Now I am not so sure...Gary says to upgrade Parity BEFORE you move, but wouldn't that keep you on RFS on Parity volume? Is the rest of this correct?
Archived
This topic is now archived and is closed to further replies.