Cache Drive issues


19 posts in this topic Last Reply

Recommended Posts

Ok so now I've royally screwed up.  I had watched Spaceinvader's video and got ansty about using a spare SSD i had laying around.

 

Here's how my server operated previously:

I had one Cache drive.  I ran a couple dockers (Plex and Valheim) off of it as well as used it for my shares to temp. store data.  Mover schedules to run every night at 3:40am.  The cache drive was simply called "cache".

 

I updated Unraid with a new Cache SSD.  I added a new pool and called this "Cache_ssd".  I renamed my previous cache drive to "Cache_nvme".   Now the issue is that I do not believe i stopped my dockers prior to renaming the cache nvme drive.  Now when I open my dockers, it's as-if I don't have any previous set ups and it want's me to set Plex and Valheim Gameserver up from scratch.   In my appdata folder; however, the files are all there.

 

How do I correct this?  I tried updating the docker.img, starting/stopping docker, etc.  I can't figure it out.   I feel like unraid still thinks theres a disk out there called  "cache"  and I just need to copy those files over to a disk called "cache_nvme".  When I open Krusader, I don't have a folder in there called "UNRAID" which from what I see in instructional videos is where you can access the disk by disk directories.  Please help as my plex library is huge and I'd rather not rebuild with all my settings, etc.  I think this error is supported by the fact I have a "Fix Common Errors" problem of the following:

 

Invalid folder cache contained within /mnt               Generally speaking, most times when other folders get created within /mnt it is a result of an improperly configured application. This error may or may not cause issues for you.  Also below is a snippet of my tree -L 2 /mnt/.  So i guess the ultimate question is how do i get the data off of "cache" and onto "cache_nvme"

 

Thanks,

Hawkins!

Capture.JPG

Edited by Hawkins12
Link to post

Ok so I partially figured out the issue.  Using Midnight Commander, I was able to see that my old stuff was in the "cache_nvme" folder and the new stuff was in the "cache" folder.  The dockers were looking out to "cache" instead of "cache_nvme" to run the data.  Once i updated dockers to look in "cache_nvme", the issue was fixed.

 

With that being said, I still have this rogue "cache" folder out there with stuff from the old docker.  How do I get rid of that?  Simply just deleting through Midnight Commander?

Link to post
15 minutes ago, 6of6 said:

I don't have an answer, but please keep posting. What is "Midnight Commander" and how does it relate to cache pools?

 

6.

Being new to Unraid (and just teaching people in general), I try to lay out my steps of problem solving so others can learn.  I can say I have solved the issue above.

 

Midnight Commander is a similar "program" to Krusader.  Essentially MC is a built in Krusader.  Some will say Krusader is fasters, others say no difference, but it's a way to move/delete/copy/etc. files from one share to another or one disk to another (don't do disk to share or vice versa - i've been told).

 

Midnight Commander is accessed by "SSH'ing into the server.  I use Putty but when you log in via SSH, you simply just type  mc  and it will open Midnight Commander.  Within Midnight Commander, I was able to find this rogue "cache" drive and delete it.  The problem has been fixed and I figured out the issue.  Essentially, my dockers were looking for a folder like "/mnt/cache/appdata"  when really the dockers needed to be lookcing for "/mnt/cache_nvme/appdata".   All my docker data (plex and valheim) were stored in "cache_nvme" however server was calling for "cache" and was just creating new files because dockers were not linked to cache_nvm.  Once updating the docker appdata info to reach for cache_nvme, dockers work as they used to and issue resolved.  I deleted "cahce" via MC and all is well.  

Link to post

Thank you!... I'm pretty much paranoid about security and have SSH disabled (a monitor and keyboard attached is the only "terminal" to my unraid). Why did Krusader not work?...

 

I'm sorry... I asked you for answers that you gave and then asked questions off topic. First two words of my post remain... Thank you!

 

6.

Link to post
52 minutes ago, 6of6 said:

Thank you!... I'm pretty much paranoid about security and have SSH disabled (a monitor and keyboard attached is the only "terminal" to my unraid). Why did Krusader not work?...

 

I'm sorry... I asked you for answers that you gave and then asked questions off topic. First two words of my post remain... Thank you!

 

6.

No problem.

 

Not a knock on one community or the other but I just converted within the last month from FreeNAS to Unraid.  I'll say the Unraid community seems much more helpful and patient.  That combined with the features and GUI/functionality of Unraid, I wish i would have never used FreeNAS.  

 

As to Krusader not working, I am not quite sure.  Even with Docker's enabled, I just could not figure out how to find information stored disk by disk.  I could find my shares but could not find the disks.  If i were able to find the disks in Krusader, then it would have worked fine (and I searched around for quite a bit of time).  

Link to post
1 hour ago, Hawkins12 said:

As to Krusader not working, I am not quite sure.  Even with Docker's enabled, I just could not figure out how to find information stored disk by disk.  I could find my shares but could not find the disks.  If i were able to find the disks in Krusader, then it would have worked fine (and I searched around for quite a bit of time).  

You would have to create a specific path in the Krusader template, I think it was explained in one of SpaceInvaderOne videos a while back. And I wonder if the next video will not cover that in his new series of basic video in the context of 6.9.

Anyway, you should get better answers in the specific support topic for the Krusader release you are using (there are 3 different repo for it).

 

8 hours ago, Hawkins12 said:

With that being said, I still have this rogue "cache" folder out there with stuff from the old docker.  How do I get rid of that?  Simply just deleting through Midnight Commander?

If it is empty or if the content is not important to you, yes.

Link to post

If you're looking for some more in-depth answers to how unRAID works, how to find your data, how to manage it etc, I'd be more than willing to give you a few solid answers in exchange for a few solidly laid out questions.

 

In my experience, there is nothing better to get Down And Dirty with the internals than to fire up ssh, log in as root, and start working.

 

I'll admit I kinda skimmed most of the middle bits, but I saw your problem and your solution and that you're still not sure how things really work? Are you familiar with Windows command line operation? Same for Linux? Do you know anything about SSH? I'm more than happy to fill in some gaps, and walk you through it, if you're willing to put effort into specificity in questions, and admitting when you don't know what I've just said.

Link to post
21 hours ago, ChatNoir said:

You would have to create a specific path in the Krusader template, I think it was explained in one of SpaceInvaderOne videos a while back. And I wonder if the next video will not cover that in his new series of basic video in the context of 6.9.

Anyway, you should get better answers in the specific support topic for the Krusader release you are using (there are 3 different repo for it).

 

If it is empty or if the content is not important to you, yes.

 Was able to delete just fine using MC.  All is working properly.  Thanks for taking the time to respond and assist!

 

20 hours ago, codefaux said:

If you're looking for some more in-depth answers to how unRAID works, how to find your data, how to manage it etc, I'd be more than willing to give you a few solid answers in exchange for a few solidly laid out questions.

 

In my experience, there is nothing better to get Down And Dirty with the internals than to fire up ssh, log in as root, and start working.

 

I'll admit I kinda skimmed most of the middle bits, but I saw your problem and your solution and that you're still not sure how things really work? Are you familiar with Windows command line operation? Same for Linux? Do you know anything about SSH? I'm more than happy to fill in some gaps, and walk you through it, if you're willing to put effort into specificity in questions, and admitting when you don't know what I've just said.

 

Thank you for helping the community.  I guess my issue is I am not a "coder" and understanding it is comparable (to me) like understanding a new language.  I am slowly getting the hang of it but it takes time and patience.  I very much appreciate your offer for Q&A.  I'll def. be thinking of a couple good ones; however, even that is difficult to me.  I often don't know what the questions are until I run into an issue.  I extensively try to research (via reddit and unraid forums among other avenues--youtube) and ultimately I think I figure it out but I struggle to understand what would be good to know at the surface.  I feel I am tech savvy but command line prompt is not something I am all too familiar with and will take a lot of practice for me to understand (ie. i see people using sudo commands a lot but don't even know what it really means).  Rather than have you explain that kind of stuff, I need to dive into it myself first to not waste anyone's time.  I think my issue is I am trying to be a "home network administrator" but that's on top of a very demanding job and family life so it makes my ability to find time to learn very difficult.  I will certainly "bookmark" your name and as I think of some good solid questions, I'll shoot your way.  I very much thank you for offering to assist.     As to answer your questions -- you are correct I am still not completely sure how things work.  I am vaguely familiar with command line operations.  I don't know much of anything about Linux and SSH except that I can use SSH to access my server (I have PuTTy) and how to get to MC.

Link to post

It really is like a new language, in a new country with different objects than you're used to. Don't stress yourself, everyone started there some time. Dig as much as you feel the need, feel free to PM me if you want me directly, but this forum is full of plenty of smart people so don't be afraid to throw a wider net.

 

Sometimes knowing the right question to ask is the most difficult part.

 

One thing I hear people stress about often is disk identification when moving hardware. The disk names may change, but it's actually fairly easy -- with the SMART data available in the UI, you can match a drive name (aka Party or Disk 2, or Cache) to a model and serial number. Absolute lowest-knowledge method from there is to just reboot adding one drive at a time connected to your new hardware, If you put in a disk that was previously Disk 3, note its name and reboot with another one. Sometimes it's even easier than that and you can just stuff them all in, reboot, and reason out the names yourself.

 

The good news is, no matter what happens, if you can't figure it out at all ever, unRAID does not use any special on-disk format; any Linux computer can read the data from each individual drive, assuming you didn't go out of your way choose encryption, and then forget your key -- so even in the worst case, your data won't be *lost*.

Link to post
6 hours ago, codefaux said:

One thing I hear people stress about often is disk identification when moving hardware. The disk names may change, but it's actually fairly easy -- with the SMART data available in the UI, you can match a drive name (aka Party or Disk 2, or Cache) to a model and serial number. Absolute lowest-knowledge method from there is to just reboot adding one drive at a time connected to your new hardware, If you put in a disk that was previously Disk 3, note its name and reboot with another one. Sometimes it's even easier than that and you can just stuff them all in, reboot, and reason out the names yourself.

 

The good news is, no matter what happens, if you can't figure it out at all ever, unRAID does not use any special on-disk format; any Linux computer can read the data from each individual drive, assuming you didn't go out of your way choose encryption, and then forget your key -- so even in the worst case, your data won't be *lost*.

 

One confirmation on what you mentioned.  If I understand you right, Unraid already captures this SMART data and there's nothing on my end to do.  I actually have the drives organized in a methodical format for drive identification purposes so I should easily be able to swap them over to the right respective spots but want to make sure there's nothing in UNRAID i need to do prior to the conversion.  

 

Thanks!

Link to post
32 minutes ago, Hawkins12 said:

 

One confirmation on what you mentioned.  If I understand you right, Unraid already captures this SMART data and there's nothing on my end to do.  I actually have the drives organized in a methodical format for drive identification purposes so I should easily be able to swap them over to the right respective spots but want to make sure there's nothing in UNRAID i need to do prior to the conversion.  

 

Thanks!

The general advice is to have a screenshot of the Main tab as if the new system identifies drives differently that should be e nough to let your sort out which drive belongs where.    Most of the time the screenshot is not needed but it is an easy precaution to take.

Link to post
1 hour ago, itimpi said:

The general advice is to have a screenshot of the Main tab as if the new system identifies drives differently that should be e nough to let your sort out which drive belongs where.    Most of the time the screenshot is not needed but it is an easy precaution to take.

Thanks.  I certainly will do this.

 

However this was not the case when I pulled the drive out accidently (I posted in another thread about this).  I accidently pulled the wrong hot swap drive the other day and reinserted it.  Grant it, the array was still running when I pulled the drive.  When i re-inserted, stopped the array, mounted to same disk spot, it indicated the drive was emulated and I had to rebuild the disk.  Was the only reason this happened due to the array being on when drive was pulled?

Link to post
17 minutes ago, Hawkins12 said:

Thanks.  I certainly will do this.

 

However this was not the case when I pulled the drive out accidently (I posted in another thread about this).  I accidently pulled the wrong hot swap drive the other day and reinserted it.  Grant it, the array was still running when I pulled the drive.  When i re-inserted, stopped the array, mounted to same disk spot, it indicated the drive was emulated and I had to rebuild the disk.  Was the only reason this happened due to the array being on when drive was pulled?

Pulling a disk with the array started will be cause a disk to be disabled and thus emulated.

 

Link to post
16 minutes ago, itimpi said:

Pulling a disk with the array started will be cause a disk to be disabled and thus emulated.

 

Thanks for this.  This is what I suspected.  So as long as I shut down the server/Unraid, I can convert over to new chassis and Unraid should recognize every disk as it was before?

Link to post
7 minutes ago, Hawkins12 said:

Thanks for this.  This is what I suspected.  So as long as I shut down the server/Unraid, I can convert over to new chassis and Unraid should recognize every disk as it was before?

 

I have done that several times in the past without issues.

 

The key thing is that if any step does not go as expected be prepared to ask for help (rather than attempt a recovery if you are not sure what is going on) so that the optimum next action can be determined.

Link to post
27 minutes ago, itimpi said:

 

I have done that several times in the past without issues.

 

The key thing is that if any step does not go as expected be prepared to ask for help (rather than attempt a recovery if you are not sure what is going on) so that the optimum next action can be determined.

Sure thing.  I'll be tackling the project this weekend so I'll be sure to ask for help if something goes awry. 

 

Thanks!!

Link to post

Wow... This thread changed. STOP!... (might not have read everything) but, Make sure you know every assignment and every serial number for every drive in your system (including the main array and every cache pool and every unassigned device). I know it sounds obvious... but so many don't do it. I have all my drive assignments backed up on another Linux box AND my Windoze box.

 

Just my two cents.

 

6.

Link to post

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.
Note: Your post will require moderator approval before it will be visible.

Guest
Reply to this topic...

×   Pasted as rich text.   Restore formatting

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.