How to Safely Shrink an Unraid Array

[@SpaceinvaderOne]

For 6.12, you'll need to either modify the script or do the actual disk clear yourself (if you do it manually don't create the clear-me folder). To modify the script, search for "md${d:9}" and replace it with "md${d:9}p1" (you have to select the partition in 6.12.x). Then all the steps should work. To just do the clear part manually, run: "dd bs=1M if=/dev/zero of=/dev/mdXp1 status=progress" (replacing x with the drive number). The full steps can be found in this forum post: forums.unraid.net/topic/61614-shrink-array-question/?do=findComment&comment=606335 (thanks Daniel K for emailing me and bringing this to my attention 🙂)

[@dchsueh]

Did this process on 6.12.3; on stopping the array I get into an endless retry loop (seen in logs) where it tries to unmount drive but reports "umount: /mnt/diskX: not mounted." I had to "reboot -n -f" to force proceed to next steps. and despite multiple errors on the reboot, a parity check succeeded.

[@39zack]

Used Notepad++ to find and replace all 5 instances of "md${d:9}" with "md${d:9}p1"and it looks to work just fine now on 6.12.4 :)

[@organom]

​@@dchsueh just got the same issue. Since I was afraid to restart unraid in this state, I created a local virtual disk using `dd if=/dev/zero of=diskX bs=1M count=310` (xfs requires at least 300mb disk), created filesystem with `mkfs -t xfs diskX` and mounted as `mount diskX /mnt/diskX`. Almost immediately after doing that, the stopping of the array was able to complete (cleanup the created virtual disks by running `rm diskX` after the array is stopped)

[@omark1781]

@@organom, you are a life-saver, this script will cause this problem on latest version of unraid. I was about to do an unclean shutdown, but your script stopped the array. @SpaceinvaderOne, please add this step in your video

[@omark1781]

can you add @dchsueh unmount step below. seems to be an issue w/latest issue of urnaid

[@JeffKogut]

Imagine how much better Unraid would be if method #2 was all in a Tool/Wizard just built-in. Just check a box of a drive to remove and all this is automated.

[@Bacender]

Yes please. So many processes in Unraid require multiple steps in a very specific order that fall apart if you have some sort of edge case where your config is not exactly like the tutorial you're trying to follow. A simple wizard with a couple clicks to begin the process would be so much better.

[@stillthinke]

This should have been implemented a long time ago, lets hope in 2023 it happens, would be amazing.

[@Kamryx]

It's actually unfathomable that this is a question you have to google and follow a video like this to resolve. Is it seriously that fringe of a use case to, *check notes*, remove a hard drive from a cluster of hard drives? Feels like I'm handling a volatile nuclear bomb for what should be 2 button clicks.

[@warham4155]

well 2024 still no @@stillthinke

[@espressomatic]

If there was a tool built in you wouldn't actually have to clear the disk at all, that part is useless. Including the necessary data in the parity to represent that disk has been zeroed is all that's necessary - not also writing to the disk being removed might save some time. At the very least you can use that disk somewhere else in the meantime, including the recycling bin if it's bad.

[@dannykempkes4957]

Another great video. Watching the video made me realise there was a third option in my case. It's basically the same as option 1, but instead of reusing the existing parity drive, I used another (pre-cleared) drive of the same size and set that one as the new parity drive. In case of failure, i would always have the original parity drive as a backup. Once the parity had been rebuild on the new disk, I upgraded a smaller data drive with the old parity drive.

[@adhatcher73]

This worked like a champ. I was replacing a 2TB drive with a 12TB drive, so zeroing out the 2TB drive was a lot faster option than rebuilding the parity drive. Thanks so much! Love all your videos, although they are a blessing and a curse. I can't tell you how many times I've watched a video and 3 hours later, had a new feature on my Unraid server. LOL.

[@MrZomhad]

Can't believe only 8000 people were lucky enough to see this, thank you so much for showing the zeroing the drive out method, saved me tons of time!

[@SteveJones172pilot]

This is a great video. All of your videos are so well done.. Hitting every detail without wasting a bunch of time.. Definitely going to have to use the zeroing out drive in the future!

[@xpgx1]

Uhh, another must-see for anyone who wants to understand his new unRAID Server *more* =) Love your videos, they are spot on - factual - funny and can be really easy to follow along! Honestly, it's a joy to learn more about unRAID this way.

[@KoncadeGaming]

On 6.11.5, used method 1 and removed 2 hard drives successfully. Everything worked as described. Only thing I did additional is once removing the data off the 2 HDDs using unblance I also removed those drives from each share drive setting then did a new configuration. Thanks very much for this tutorial, was sh1tting bricks nervously until parity did the sync successfully. I agree that this should definitely be a wizard within unraid to do all this automatically.

[@Marc-vr5vg]

Great instruction, thank you so much! I was able to prepare my move from 5 HDDs (2x 14, 2x4 & 1x6TB) to an array of 3x16TB by removing 2 drives using your method. When all my hardware arrives I will be replacing the disks 1 by 1 to new ones. Great stuff! I used the 2nd method by the way.

[@homesteadishdad]

Thank you so much for this video. I've been trying to fix my array to no avail and this has been completely game-changing. By minute 24, you had helped me and I appreciate you ending my frustration!

[@bluedevil678]

I cant thank you enough for these videos - you have exspanded my and many others minds with these guides. A massive thank you!

[@tychamberlain7125]

Seriously, I literally needed this video two days ago! Haha. Keep up the great content!

[@pakcheesy2]

Same

[@aliengoa0]

greatest videos of all! Bought an Unraid license and mentioned you in the form

[@henkeenor]

Thanks a lot for a great tutorial (as always)! I’ve just followed this method to the dot to zero and remove a 4TB drive with parity intact. No hiccups, took about 10,5 hours (avg. 103 MB/s).

[@lightwiz267]

Thank You so much for this video. I have a dead drive in my server and I posted this question on unraid just last week. Didn't get any descent answers. Knew there had to be a better solution then copying all the data off.

[@LordShad0w.162]

Thanks again Ed! Just had to do this 5 minutes ago after a cold spare was found to be faulty AFTER a full clear operation and smart eval. :(

[@slstevens5475]

Ed, as always a great tutorial. How about doing an update based upon UNRAID updates, lessons-learned since your original video was made, and user questions and comments? Thanks for all you do!!

[@droknron]

Thank you so much for this video. I was able to follow the method 2 and maintain parity while removing a data drive thanks to it. Very thorough and informative :)

[@descendfromgrace1]

Thanks so much for all that you do for the community! This video helped me a lot.

[@subukai]

Thank you for this guide. I was following the documented process on the Unraid site but this video gave additional information not covered. I noticed the clear disk method 2 started at 15:39 and ended at 18:03 for a 1 TB clear. I wanted to get an idea of how long this clear is going to take. So my 4 Tb drive theoretically will take 9+ hours. Thanks!

[@DanKaschel]

I am about to remove a 4TB drive -- did it take about 9hrs to clear for you?

[@subukai]

@@DanKaschel pretty sure it was that long but that was a year ago.

[@wilsmith7173]

Followed instructions for Method 2 without a hitch. Can recommend!

[@cmiller6352]

How long was the process to clear the drive? And how large was it? Sitting at 12+ hours for a 1TB clear over here. Seems long, but the process is active so it hasn't stalled.

[@wilsmith7173]

@@cmiller6352 mine was 6tb and took about 16 hours. Happy with result.

[@cmiller6352]

@@wilsmith7173 As expected, within minutes of me asking, the clear finished. haha Thanks for the reply.

[@ScottBalkum]

Thank you for your vids. They are very detailed and very helpful on my journey of Unraid.

[@sixtyfps]

You are a marvelous teacher. Thank you for helping me put the shine on my formerly deprecated (reiserfs???!) Unraid setup!

[@erichankins2634]

Thanks again sir!! you and SPX Labs are the best!

[@chironjit]

A note of warning that the latest UnRaid version breaks this somewhere for the second method(most likely the script that is meant to zero the drives). I tried it with mine and it essentially had to rewrite the whole parity again. Would suggest running a parity check once you've completed all the steps

[@marshallhoward7937]

Ironic we're doing the same thing. For me, the disk clear script finished almost immediately & the disk shows virtually no used space. However, Trying to stop the array it just hangs. Rebooting shows space back on the drive. Guess it's method 1 for me, rebuilding parity. Sucks, I have 8 2TB drives that i need to do this to.

[@chironjit]

@@marshallhoward7937 that is EXACTLY what happened to mine too! The script fails to clear the disk and just leaves it at some unmounted state. UNRAID can't unmount the drives because the drives are already unmounted but also refuses to mount then again, so i end up having to force a reboot. Unfortunately mine are 10TB drives, so its literally 36 hours for a parity check

[@chironjit]

Will report back on whether my data is corrupted once the parity rewrites are done. I just followed the rest of the steps to method 2 and ran a parity check once i started the array with the drives removed

[@omark1781]

Any issues after parity check?​@@chironjit

[@omark1781]

So after running the script in option 2, everything went well but when i tried to stop the array, it would hang on unmounting the drives. Any fixes/suggestions for this?

[@Goldmaster]

23:05 what about drives that are now labelled as disk 1, then disk 3? Any way to make disk 3 become disk 2, without making anything more complex?

[@MrMischelito]

I don't think so. Does it bother you?

[@TodPalin]

@@MrMischelito Yes

[@sih8283]

​@@MrMischelito Yes

[@romankiss3793]

Thank You, A Video how to use a second unraid server for backup would be great

[@IEnjoyCreatingVideos]

Great video as always Ed! Thank you for sharing it with us!💖👍😎JP

[@wgolden3]

Great video. Exactly what I needed

[@MarvelJAM]

I cannot tell you how many times I have come back to this video as a reminder. Great one. Thanks for the content!

[@DarkLinkArt]

Thank you for all the great tutorials!!!

[@mrq332]

love your video's, made my unraid server like i want it based on your video's, thanks!

[@playati24]

I've been waiting for this! Amazingly useful, as always. How long does the clear disk typically take for larger drives (14-16) for the retained parity method? Same as a parity check?

[@GodmanchesterGoblin]

If the drive being cleared is as large as the parity drive, then yes, the time is about the same. But if you are removing a much smaller drive, then it will only take as long as the parity check would take before it finishes checking the drives of that smaller size. If you look at the disk access stats during a parity check, you can see a series of steps in the disk throughput for each drive size that you're using. This would also give you an idea of how long it takes to clear a smaller drive.

[@playati24]

​@@GodmanchesterGoblin​Thanks - super helpful. That's a pretty significant amount of downtime (~25-30hrs), but I suppose that's probably the right decision compared to rolling the dice with a parity rebuild.

[@GodmanchesterGoblin]

@@playati24 Personally, I never take those chances, even though I have a second server that is a 100% backup of the main one. Probably it's just my approach to the satisfaction that I get from knowing that my data integrity is assured if I follow a particular set of options, but it has never let me down. I have files from the late 80s and early 90s - I would just hate to lose that stuff, even though I could live perfectly well without it.

[@neighborhoodtechgeek2954]

OMG this is amazing. Didn't know this existed. :D

[@maxdata8861]

First I want to say I love your videos! Many have saved me hours I probably would have wasted!. Any chance you are considering making a video on fixing the "low current" startup that causes the dreaded DISABLED condition with new Seagate drives and Unraid 6.9.x ? I have had two new fail in the past month. Thanks again for you r efforts.

[@Gemini1st]

Thank you a lot. Very useful video 🙂

[@harikili]

Thank you for this video.

[@sidarthus89]

So i got to 11:30 and say the data with the parity drive in the unassigned devices. Went to go stop the array again to add the parity drive. But its hanidng at stopping the array. Stating Retry unmounting disk shares. That disk1 target is busy. Am i safe to do a normal shutdown at this stage?

[@glasshalfempty1984]

14:11 what would be the advantages, if any, of having specific content on only certain drives?

[@Andrew-em6mt]

For me, having one specific tv show with all its seasons on the same disk means I only need to use the life of 1 drive. If split across multiple drives, I would need to spin up multiple drives thus consuming the life of multiple drives. For example: I have a 4 disks which I spread my tv shows across, however the episodes of any given tv show remain together on the same disk. You will not find 1 tv show which has its episodes split across multiple disks. If i get to a point where that disk is full, and need to add more episodes to a particular tv show, those episodes won't get moved. I would need to move the whole tv show to another disk that has more space for mover to work again. If you watch as many tv show episodes as I do, its worth doing 😜

[@nl4771]

Thanks. I was just about to do this with my server.

[@sublimnl.]

Crap like this is why I am literally moving off of unraid right now. Any time you have to touch the array it's a tedious process fraught with red flags and danger. Trying Xpenology now, which will hopefully be a lot more hands off and easier to manage when I do need to make changes.

[@turkeyphant]

Can you move disk 3 back to disk 2 so there are no gaps after finishing method 2?

[@FolixOrision]

This was helpful, thank you

[@DarksurfX]

clear drive script is broken :( it doesn't appear to work on 6.10.3 and immediately gives up and claims drives are not empty even though they ARE empty (I've personally verified that myself), and the "clear-me" directory is there.

[@scoops2]

Something may be wrong with the script on latest unraid 6.12.3 in August 2023 I followed the steps exactly and the script ran way too quickly but the logs were exactly as shown in the video. After that my array was no longer able to stop because it could not spin down that disk anymore. I decided I'd just rebuild parity so I shutdown the server and physically removed the disk, now even with that disk removed other disks are "busy" and can not unmount down. I'm now troubleshooting to see what can be done.

[@AlienShine]

Tedious process. would be nice if it worked for multiple drives at one. Replacing 3 4tb drives is not fun. EDIT: just followed you first method with the parity rebuild, seems the better option when replacing multiple drives, so thanks for showing both ways!

[@theericgauch]

Danke!

[@dalboy1506]

Do these methods still work? Or is there another way to shrink the array while maintaining parity? I'm suffering the same issue as others have mentioned that the script reports that the disk is not empty, even though it is.

[@antpk00]

Just a double check but method 2 wouldn't work with an encrypted xfs formatted drive. Am I on the right track?

[@jasonmuller2225]

If I use method 2, can I just keep the Pool slots checked when doing a new configuration or do I need to click ALL?

[@MrMischelito]

Excellent procedure

[@glasshalfempty1984]

15:28 I'm very curious about this situation. It's a backup server, so wouldn't you want those backups to be protected with extra redundancy? What's the point of a backup server at all if you use two parity drives and backup your file server content off-site? (This is my situation) Just trying to understand this use case scenario.

[@activate__motivation]

[@WindowsVistaFTW]

FYI this zeroing script from method no longer works as of 6.12. The command to zero now requires the partition # to be specified and the script hasn't been updated.

[@jonirish323]

I just tried to do this with my UnRAID 6.12.2 server, and unBALANCE will not run. I turned off Docker and VM services and changed the schedule on mover, but from "Scatter" all 3 options: PLAN, MOVE, and COPY all display the circle with a hash through it and refuses to run. I noticed that unBALANCE doesn't appear to have been updated since the v6.12 release. Is it possible that it isn't compatible with 6.12? If so, is there any other way tor remove unused drives from the array w/o having to rebuild parity?

[@Andrew-em6mt]

This is super weird. So I moved all my data that was stored on an old 2tb drive. anyway, after a restart, one of my 10tb drives which is only 2 months old (and passed preclear tests) become disabled due to having almost 11,000 udma crc error counts (suspect bad slot in my caddy and/or HBA card. now i went ahead and followed method 2 above by taking a chance as i needed room to expand plus replace the suspected failed drive by putting the new one in the same slot as the 2tb (so i can isolate the HBA cables and the slot in my caddy). After running the script, turning off the array and removing the 2tb drive from the array, the 10tb is now no longer disabled and working as normal!! would that have anything to do with the "new config" steps?

[@theram8653]

I made the mistake of forgetting to change the tunable MD write method before running the script. How will this affect me? The script has been running for about 12 hours now which is longer than normal (i have done this process about 3 times total, each time removing a 4tb drive)

[@turkeyphant]

Does this work with dual parity?

[@miicar204]

Another great walkthrough...although i just did this on 6.12.2 and it hung trying to unmount on the newly "cleared" disk trying to stop the array. so i turned off auto start, and rebooted to remove the drive.

[@stillsb]

perfect and big love！

[@espressomatic]

If you follow the instructions from Unraid's Wiki and format the drive you intend to remove, the script may fail because it'll think the drive isn't empty. Example, format btrfs and the "du -s" in the script returns 16 - anything non-0 causes failure. Formatting for a second time, back to xfs, seems to solve this.

[@cptfbi]

Can I do this clear-me if I have 3 disks to remove at once?

[@tedroddy]

you are a god amongst men

[@HacksawJimThuggin]

Once you remove the drives, is it safe to move them around? Example, can you move disk3 to slot 2?

[@neilmandalaywala9991]

This is perfect! I just bought a bunch of 18tb drives for my server to replace my 8tb drives in my node 804. Thank you for this. Question: I have a second I raid server. I was using it as a virtual gaming rig but ended up ripping the card for an actual PC. I want to take these old 8tb drives (10 of them) and now make a backup of my primary server which does pretty much everything I want it to. Is there a program you use to setup backups like you discussed in this video?

[@joelang6126]

Simple rsync script with the second server mounted will do.

[@chrislowe8085]

Hey as of 6.12.3 the drive clean script does not work. i was forced to say new config and as i mounted the array i checked the parity is good all green. is there a fix for the drive clean script? thanks for the help with all your unraid video's there fantanic

[@chrisblair1634]

RUN SCRIPT IN BACKGROUND!! If you don’t run the script in the background and accidentally close the terminal window or even the browser window, the script will stop and it will cause problems. ONE DISK AT A TIME!! If you’re shrinking your array by removing multiple disks, save yourself some headaches and do one disk at a time. ‘CLEAR DISK’ SCRIPT TAKES 1.5 HRS PER TERABYTE!! Expect to wait 12 hours or more for the script to write zeros to the disk.

[@Andrew-em6mt]

That is odd? Maybe your drive was really old, wrecked or did you forget to change the Tunable md write method setting? In his video, the log shows it took 2.5 hours to complete 8TB.

[@miicar204]

Does unbalance support multiple cache pools yet? i had some major issues happen when i tried using it with multiple pools....BEWARE (use Krusader or R-sync)

[@misza_one]

Can this be done the same way with encryption enabled? Or is there something else to consider?

[@MrMischelito]

Should work just fine

[@dieseldes6578]

In October 24, is this still the best method to follow? Thanks

[@johngessner]

First thanks for the video. I did this for three small drives and I guess I should have removed each one as it finished, as after all that my server still wanted to do a parity check. I did a new config at the very end to get rid of the missing drives that stilled showed. Maybe that was the issue with the server wanting a parity check.

[@patricklayton9748]

Doesnt work for me getting a closing socket: unable to ReadMessag error

[@mobi8138]

I tried using the script method and it would prevent me from shutting down the array. The script would unmount the drive and leave it unmounted so you couldnt stop the array.

[@BillOstrowski]

Thanks for this video. I do have a question. How do I remove a cache drive? Is it the same process and don't save the pool?

[@jeffhynes4418]

mine is only going about 5 MB/s - any ideas why?

[@SteveJones172pilot]

Why is it important not to use the array at all while a drive is being zero'd out? I know it could be a performance hit, but as long as the drive being cleared isn't written to, is it really a bad thing?

[@IzzNB]

Why is it that every time I try this, things go great until I try to stop the array? It just keeps trying to unmount the disk that was just cleared. Unmounting disks... umount: /mnt/disk1: not mounted. shcmd (960): exit status: 32 Retry unmounting disk share(s)... I end up having to force the system to shut down to continue.

[@omark1781]

same thing happened to me, looking for a solution because i have 2 more disks i have to remove. might just risk it and go with option 1

[@ahwayzcool4630]

I just built an unraid system a few days ago and i have two new 14tb drives with no data on them id like to remove. I take it i need to still go thru this process correct? Even if i have no data that need to be moved from the drives?

[@SpaceinvaderOne]

Great question! When you delete data, it doesn't always get overwritten with zeros, which means there may still be some leftover data on the drives if you have written data to these drives then deleted it. If you were to simply remove the drives without properly wiping them, this could cause issues with your parity data. So, it's highly recommended to follow the proper process of removing the drives from your Unraid system. This will ensure that all data is completely erased and the disks are properly zeroed out before you physically remove them from your system. This way, your parity data won't be affected by the removal of the disk. Hope that helps! Let me know if you have any other questions or concerns.

[@ahwayzcool4630]

@@SpaceinvaderOne Ok thanks for your response. I ended up following your tutorial a few days ago and successfully zeroed and removed 1 drive. Working on zeroing a second drive but I forgot to change that one option in disk management that speeds up the zeroing process and now its taking FOREVER. Smh.

[@ahwayzcool4630]

@@SpaceinvaderOne One last question. I was looking for your second video tutorial on making a backup server for your main unraid server. Is there a link for the video? There's not many tutorial videos out their on doing this and I rather watch your tutorial because you explain things in great detail. 🙂

[@kesslerdan1]

Can I use the 2nd method (disk zeroing) to remove 2 disks at the same time by running the script on the 2 disks simulataneously (creating a "clear-me" folder on both after empting them ofcourse)?

[@shawnng5824]

Does it work for you? 2nd method, 2 disks simulataneously?

[@matthewkrupin8195]

(a few months later..) what you can do is make duplicates of the script N times for the number of drives you want to clear, and update the line of code that says "marker=clear-me" to something like "marker=clear-me-5" and rename the folder in your drive to match ex. disk 5 => /mnt/disk5/clear-me-5 the script has no knowledge of which drive is busy, so each drive and script must have a distinct folder / marker name or else the scripts will always target the same drive .. at least thats my guess, i haven't tested it so i have no idea what errors would occur if the script ran multiple times in parallel for a single drive i'm currently doing these steps for 3 drives and it seems to be working fine note however there's no guarantee of a speedup because the process can and will be throttled by the parity drives

[@vanvee88]

Could option #1 be done if I wanted to shrink 3 disk to 1? If not what is recommended? I'm fully backed up so fine with parity checks and uptime isn't a priority. Spaceinvader One is amazing.

[@MMorris75]

Script has ran for over 8 hours read and writes have stopped showing on parity, but now several drives including parity are showing errors. Has anyone else seen this? I'm on 6.10.

[@optimustrond]

Thanks for another great guide! Don't know what I would do with out your videos! I removed one drive using the preserve parity method, but I still got 13 sync errors corrected after I did a parity check. Any idea what could have gone wrong? I stopped both the dockers and the vm's, and as far as I know, nothing has written anything to the array. Might have been a fluke, but I am doing another parity check now just to double check that I'm not getting any errors now.

[@tmatzdorf]

I tried doing this. 3tb and 3 days later its still running.

[@jirayah5435]

can you do method #1 on multiple drives and zero all of them at the same time?

[@joelang6126]

Ed, I assume you've given up on the final to Tdarr video comparing GPU/CPU encodes?

[@ForthpalmUK]

Sadly the script does not appear to work in 6.12.3 - "Unmounting Disk 6 ... umount: /mnt/disk6: not mounted. Clearing Disk 6 ... dd: error writing '/dev/md6': No space left on device 9+0 records in 8+0 records out 8388608 bytes (8.4 MB, 8.0 MiB) copied, 0.00389879 s, 2.2 GB/s" that's a 6Tb sata drive. Script used to work great but no longer it would appear :(

[@kdolanjr]

I did this incorrectly just 4 days ago but fortunately xfs repair saved my butt. Removed drives had no data left on them, but the file system wasn't happy 😬

[@Meerlu]

You should put a warning about how long zeroing a drive takes .. its friday and mines been zeroing since... MONDAY evening!!!!!!

[@K1LLA_KING_KONG]

Is there a modern update to this method? Like a built in tool?

[@DanKaschel]

It's important to remember that this is a server, and this *is* a modern method for a server. Many are asking/looking for a consumer/novice-level tool to accomplish this, which doesn't exist to my knowledge. That is the hazard of running unraid--although most tasks ARE quite easy and GUI-driven, you will occasionally have to deal with the fact that, beneath the nice coat of paint, it's simply a linux server.

[@acraen]

I followed your guide.. yet something must have gone wrong. I moved using unbalance. and checked that the drive was empty I copied the script from the url provided. I executed the script (I must say it finished quite quickly (minutes for a 20tb disk?) I saved the config stopped the array and unpopulated the drive from the array. started the array indicating parity was intact. I performed a parity check to be sure parity was intact. yet I see the number of corrections being in the hundreds of millions Am I missing something? (I still have to shrink 5 disks in order to move to zfs... so would like to know how this can be prevented)

[@nprs2002]

I also had the same issue. Not sure what I did wrong.

[@shortfred2727]

Thank you for the Video! Perfect. How i can change the cache? From Raid 10 Sata to Raid 1 nvme?

[@39zack]

Hmmm, what is the benefits of running ZFS on 1 drive?

[@almulder10]

@spaceinvader One, Could you do a video about parity errors and how to find the cause and fix them. I get about 500+ each month and unable find the cause. all drives are in a DS4246 connecter to server via LSI Controller.

[@GodmanchesterGoblin]

Ouch! Something going wrong there. In a healthy system you should never have any parity errors reported. Is the number the same each time? Are the errors always on the same drive(s)? Is any of your cabling to the drives tightly bundled? (It should not be since that can give rise to errors). What does the SMART data for each of your drives look like (DMA CRC errors? Reallocated sectors? etc...)? Since these errors first started. have you ever run a correcting check to ensure that parity is correct? Note: Do NOT do this unless you are sure that the errors are real errors on the parity disk, otherwise you may create more problems that you solve. If you are unclear on what to do, I recommend that you reach out for help on the Unraid forums where people can guide you. The first step will be to describe the system and the problem and also to post your system diagnostics along with your query.

[@almulder10]

@@GodmanchesterGoblin There are no error on the disks, they just come up as parity errors, cables, lsi card, and disk shelf have all been replaced, issue still continues. Also all drives including parity are located in one diskshelf DS4246. I have the done the error correcting check and not the error amount is different each month, but seems to be between 400-500 range. All drives have been checked, there are no errors, nothing in SMART. this is why i am so puzzled.

[@GodmanchesterGoblin]

@@almulder10 Weird. I used to have a card that consistently gave just 5 errors - that was a Supermicro AOC-SASLP-MV8. I replaced with a AOC-SAS2LP-MV8 version and I have never seen the issues since. That was a strange one - the errors were always exactly the same each time. But as you say, you have changed the card. I have no knowledge or experience of the Disk Shelf products, but it would seem unlikely that would be responsible. And if you are a seasoned user then I guess you have already been on the forum.

[@radian2323]

No dice on method 2. No matter what I do it does not find an empty disk that is marked. Went in to MC to delete all files, checked for hidden files/folders, ran a rm -r /mnt/disk3/* and still no dice. Of course I added the clear-me folder after doing all this. Even restarted unraid.

[@boyturtleuk]

I had this exact same issue. Did you get it resolved, or did you go with option 1? I tried it on v6.9.2 and then updated to 6.10.3 to no avail.

[@zoltrance5051]

Why didn't you choose "all" at 10mins under preserve current assignment? Then all you have to do is choose the file system? I've been learning this whole converting rfs to zfs fiasco, had to do 5 rfs drives and it's a hf minefield! haha

[@sarf9069]

Not sure where to post this so thought I would just ask in the latest video, hope you don't mind. What I would like to do is run a windows VM from within a linux VM on unraid for certain games and apps that simply wont run via wine due to anti cheat. I fully understand I can just switch to a windows VM and this works perfectly but as my main daily driver is Linux I am using windows less and less. Maybe there is a way to do this similar to how you can run nested VM's on vmware but the solution I'm looking for would need to utilise a single GPU for the "nested windows VM" while it was running for any games. not sure if this is possible so thought I would ask an expert. Thank's in advance and love your videos, you are awesome!

[@vincentlaw1016]

Regarding method 1, you can save time by skipping unbalance and instead mount the removed disk as unassigned device and rsync the data directly into the array.

[@warham4155]

Dont use this script on modern versions of Unraid unless you really know what your doing. It will hang when trying to shutdown the array.

[@Meerlu]

The zeroing of a drive takes a lonoooooooooooooooooog time.. im over 30 hours now on a 6TB :/

[@MMorris75]

Did the drive finally zero itself? I have a 4tb drive and its been over 12 hours. Rebuilding parity wouldn't have taken this long.

[@Meerlu]

@@MMorris75 I manually stopped it after 3 days. LOL

[@gillesdavid7978]

I must be doing something wrong, I tried the 1st method and everything worked well but every time I start my server the array doesn't come online automaticallyand I have to login and push start which is difficult when your kids want to watch something....

[@gillesdavid7978]

doesn't matter found it...