Mover Running, stuck...half of cache moved..

[wickedathletes]

I am in the process of moving my data off my cache to put on a new cache drive. the mover appears stuck, it hasnt written to log in over an hour

 

Dec 8 13:47:33 Hades root: cd+++++++++ apps/Plex/Library/Application Support/Plex Media Server/Media/localhost/a/da0557cb221d85715fb47d468e8bd5cdbb0ac65.bundle/Contents/Thumbnails/

 

---that folder has 1 image in it.

 

I was using the instructions here: https://lime-technology.com/wiki/index.php/Replace_A_Cache_Drive

 

I do have an app folder backup from last night, so not sure how I can kill this process? if I shouldnt or what? It is writing to my drive with space on it, so not sure why it just died out. any help appreciated, my server is dead in the water as we speak.

[wickedathletes]

this is all my syslog says:

 

Fatal error: Allowed memory size of 134217728 bytes exhausted (tried to allocate 72 bytes) in /usr/local/emhttp/plugins/dynamix/include/DefaultPageLayout.php(300) : eval()'d code on line 73

[Squid]

this is all my syslog says:

 

Fatal error: Allowed memory size of 134217728 bytes exhausted (tried to allocate 72 bytes) in /usr/local/emhttp/plugins/dynamix/include/DefaultPageLayout.php(300) : eval()'d code on line 73

Sure that's in your syslog and not on one of the UI pages?  If so, which page?

[wickedathletes]

this is all my syslog says:

 

Fatal error: Allowed memory size of 134217728 bytes exhausted (tried to allocate 72 bytes) in /usr/local/emhttp/plugins/dynamix/include/DefaultPageLayout.php(300) : eval()'d code on line 73

Sure that's in your syslog and not on one of the UI pages?  If so, which page?

 

tools->system log.

 

screen shot: https://drive.google.com/open?id=0B7baFOqwwO97dVMwVkZiRHRLU3M

 

I think at this point I just want to know how to kill a Mover job? If I reboot will it kill it or will it just start up again?

[Squid]

this is all my syslog says:

 

Fatal error: Allowed memory size of 134217728 bytes exhausted (tried to allocate 72 bytes) in /usr/local/emhttp/plugins/dynamix/include/DefaultPageLayout.php(300) : eval()'d code on line 73

Sure that's in your syslog and not on one of the UI pages?  If so, which page?

 

tools->system log.

 

screen shot: https://drive.google.com/open?id=0B7baFOqwwO97dVMwVkZiRHRLU3M

 

I think at this point I just want to know how to kill a Mover job? If I reboot will it kill it or will it just start up again?

It's a UI page that's doing it.  If you reboot (and it works from the UI), then mover will not start back up again.

 

I know what the error is (a PHP script attempted to allocate more memory than it was allowed to) and your only real recourse is to reboot.

 

What I think happened here is that you have a sizeable Plex library, and mover logs everything which basically means that your syslog wound up with an extra 200,000+ lines in it and.

 

Since your already half way done (hopefully), just start mover back up after the system reboots.

 

 

Side Note: Using CA Backup to handle copying the appdata would have been

- Far, far faster process as it doesn't have the overhead that mover does

- You wind up with a backup copy (the original cache drive) just in case something goes wrong

- Doesn't log all moves into the syslog for this very reason (logs it elsewhere)

 

[wickedathletes]

this is all my syslog says:

 

Fatal error: Allowed memory size of 134217728 bytes exhausted (tried to allocate 72 bytes) in /usr/local/emhttp/plugins/dynamix/include/DefaultPageLayout.php(300) : eval()'d code on line 73

Sure that's in your syslog and not on one of the UI pages?  If so, which page?

 

tools->system log.

 

screen shot: https://drive.google.com/open?id=0B7baFOqwwO97dVMwVkZiRHRLU3M

 

I think at this point I just want to know how to kill a Mover job? If I reboot will it kill it or will it just start up again?

It's a UI page that's doing it.  If you reboot (and it works from the UI), then mover will not start back up again.

 

I know what the error is (a PHP script attempted to allocate more memory than it was allowed to) and your only real recourse is to reboot.

 

What I think happened here is that you have a sizeable Plex library, and mover logs everything which basically means that your syslog wound up with an extra 200,000+ lines in it and.

 

Since your already half way done (hopefully), just start mover back up after the system reboots.

 

 

Side Note: Using CA Backup to handle copying the appdata would have been

- Far, far faster process as it doesn't have the overhead that mover does

- You wind up with a backup copy (the original cache drive) just in case something goes wrong

- Doesn't log all moves into the syslog for this very reason (logs it elsewhere)

 

I used CA Backup and forgot my backup runs Wednesday AM so I have a nice clean fresh backup, which is why I don't care about mover at this point and ya, my plex library is kind of large

 

So since I have a backup point can I just plugin my new cache drive, format it and restore?

 

 

[Squid]

So long as there's nothing else on the cache drive. 

 

Sent from my LG-D852 using Tapatalk

 

 

[wickedathletes]

So long as there's nothing else on the cache drive. 

 

Sent from my LG-D852 using Tapatalk

 

cool, ya i have everything other than appdata backed up elsewhere as well.

[wickedathletes]

So long as there's nothing else on the cache drive. 

 

Sent from my LG-D852 using Tapatalk

 

Everything is up and running now, thank you. Quick one:

 

Because mover is like half finished now I am getting this error in Fix Common Problems:

 

"Share apps set to cache-only, but files / folders exist on the array"

 

If I just delete what is on my array will that go away? Is there a way to tell mover to stop thinking its 50% done in a process? All of my files are on the cache drive now.

[Squid]

So long as there's nothing else on the cache drive. 

 

Sent from my LG-D852 using Tapatalk

 

Everything is up and running now, thank you. Quick one:

 

Because mover is like half finished now I am getting this error in Fix Common Problems:

 

"Share apps set to cache-only, but files / folders exist on the array"

 

If I just delete what is on my array will that go away? Is there a way to tell mover to stop thinking its 50% done in a process? All of my files are on the cache drive now.

This is actually a weird little edge case and I'm not 100% sure how the user system handles copying to a cache-only share when an identical file exists on the array.

 

This is what I would do to make sure everything is ok

 

The apps folder sitting on the array is going to be from mover yesterday, and will be safe to kill.  Either use mc or the dolphin/krusader apps to kill the folder from /mnt/user0 or

rm -rf /mnt/user0/apps

 

Note that you must use user0 and not user to only delete the files on the array.

 

Then do the restore again.  (Only files which are missing from the cache drive version will get copied)

 

I'll have to play around tonight to see exactly what CA Backup does in this edge-case, as I know its not something that I ever tested.

[JonathanM]

"Share apps set to cache-only, but files / folders exist on the array"

 

If I just delete what is on my array will that go away? Is there a way to tell mover to stop thinking its 50% done in a process? All of my files are on the cache drive now.

As an experiment you could set appdata to "cache prefer" and run the mover again. It will take just as long, and you may still have to follow squid's directions, but it would be an interesting experiment. I suspect it will fail, but the resulting log file and summary of actions leading up to this could be useful for limetech.

[wickedathletes]

So long as there's nothing else on the cache drive. 

 

Sent from my LG-D852 using Tapatalk

 

Everything is up and running now, thank you. Quick one:

 

Because mover is like half finished now I am getting this error in Fix Common Problems:

 

"Share apps set to cache-only, but files / folders exist on the array"

 

If I just delete what is on my array will that go away? Is there a way to tell mover to stop thinking its 50% done in a process? All of my files are on the cache drive now.

This is actually a weird little edge case and I'm not 100% sure how the user system handles copying to a cache-only share when an identical file exists on the array.

 

This is what I would do to make sure everything is ok

 

The apps folder sitting on the array is going to be from mover yesterday, and will be safe to kill.  Either use mc or the dolphin/krusader apps to kill the folder from /mnt/user0 or

rm -rf /mnt/user0/apps

 

Note that you must use user0 and not user to only delete the files on the array.

 

Then do the restore again.  (Only files which are missing from the cache drive version will get copied)

 

I'll have to play around tonight to see exactly what CA Backup does in this edge-case, as I know its not something that I ever tested.

 

if i do the restore again but have changed this since then is that a problem?

[wickedathletes]

"Share apps set to cache-only, but files / folders exist on the array"

 

If I just delete what is on my array will that go away? Is there a way to tell mover to stop thinking its 50% done in a process? All of my files are on the cache drive now.

As an experiment you could set appdata to "cache prefer" and run the mover again. It will take just as long, and you may still have to follow squid's directions, but it would be an interesting experiment. I suspect it will fail, but the resulting log file and summary of actions leading up to this could be useful for limetech.

 

I am all for testing but what happens if I nuke something in the mean time? Just use the CA backup and restore again? I don't want to hose my server, its my production box so to speak.

[Squid]

 

if i do the restore again but have changed this since then is that a problem?

The restore will set your appdata back up to how it was when the backup was performed.  Any new files will get deleted, changed files will be reverted, etc

[Squid]

"Share apps set to cache-only, but files / folders exist on the array"

 

If I just delete what is on my array will that go away? Is there a way to tell mover to stop thinking its 50% done in a process? All of my files are on the cache drive now.

As an experiment you could set appdata to "cache prefer" and run the mover again. It will take just as long, and you may still have to follow squid's directions, but it would be an interesting experiment. I suspect it will fail, but the resulting log file and summary of actions leading up to this could be useful for limetech.

 

I am all for testing but what happens if I nuke something in the mean time? Just use the CA backup and restore again? I don't want to hose my server, its my production box so to speak.

Doing it through mover will take forever due to how it operates, and then you will also have the problem that the appdata won't be in sync with how its supposed to be (since the backup ran the day before you did the mover thing) - May not be a problem, but you could wind up in a situation of file A references file B, but B isn't there because mover never touched it and it wasn't in the backup set from the day prior.

 

Personally, I wouldn't bother with the test.  But you do have the backup set just in case.

[wickedathletes]

Doing it through mover will take forever due to how it operates, and then you will also have the problem that the appdata won't be in sync with how its supposed to be (since the backup ran the day before you did the mover thing) - May not be a problem, but you could wind up in a situation of file A references file B, but B isn't there because mover never touched it and it wasn't in the backup set from the day prior.

 

Personally, I wouldn't bother with the test.  But you do have the backup set just in case.

 

Is this something that could wait until the next backup and then restore from that?

[Squid]

Doing it through mover will take forever due to how it operates, and then you will also have the problem that the appdata won't be in sync with how its supposed to be (since the backup ran the day before you did the mover thing) - May not be a problem, but you could wind up in a situation of file A references file B, but B isn't there because mover never touched it and it wasn't in the backup set from the day prior.

 

Personally, I wouldn't bother with the test.  But you do have the backup set just in case.

 

Is this something that could wait until the next backup and then restore from that?

I think the key question is here is how are your apps referencing their config folders.  /mnt/cache/apps or /mnt/user/apps

[wickedathletes]

I think the key question is here is how are your apps referencing their config folders.  /mnt/cache/apps or /mnt/user/apps

 

/mnt/cache/apps

 

Has to be, mover only managed to move half of my Plex app folder before dying, everything else is working fine.

 

Is there a way to verify Plex is correct? I did have a lot of images that seemed "broke."

 

or its kind of not... crap. I see my downloads are going to my disk5 now... ugh...

[Squid]

I think the key question is here is how are your apps referencing their config folders.  /mnt/cache/apps or /mnt/user/apps

 

/mnt/cache/apps

 

Has to be, mover only managed to move half of my Plex app folder before dying, everything else is working fine.

 

Is there a way to verify Plex is correct? I did have a lot of images that seemed "broke."

Doesn't have to be /mnt/cache on the templates.  /mnt/user works just fine.

 

Best guess on the broken images is that plex is referencing /mnt/user (you can verify by editing the container) and unRaid is grabbing the version of the file(s) on the array and missing the symlinks because only half the share got copied.

 

Back to what I originally suggested.  Delete /mnt/user0/apps and restore again or even better, delete /mnt/user/apps and restore again.  Either way you'll be back to where you where a couple of days ago.

[wickedathletes]

I think the key question is here is how are your apps referencing their config folders.  /mnt/cache/apps or /mnt/user/apps

 

/mnt/cache/apps

 

Has to be, mover only managed to move half of my Plex app folder before dying, everything else is working fine.

 

Is there a way to verify Plex is correct? I did have a lot of images that seemed "broke."

Doesn't have to be /mnt/cache on the templates.  /mnt/user works just fine.

 

Best guess on the broken images is that plex is referencing /mnt/user (you can verify by editing the container) and unRaid is grabbing the version of the file(s) on the array and missing the symlinks because only half the share got copied.

 

Back to what I originally suggested.  Delete /mnt/user0/apps and restore again or even better, delete /mnt/user/apps and restore again.  Either way you'll be back to where you where a couple of days ago.

 

awesome, I think I will give that a whirl in a day or so... because when sh*t hits the fan, it hits the fan and of course I had a drive die at the same time, so right now I am preclearing and adding a replacement in the mix, so I definitely dont want to do anything right now.

[Squid]

I think the key question is here is how are your apps referencing their config folders.  /mnt/cache/apps or /mnt/user/apps

 

/mnt/cache/apps

 

Has to be, mover only managed to move half of my Plex app folder before dying, everything else is working fine.

 

Is there a way to verify Plex is correct? I did have a lot of images that seemed "broke."

Doesn't have to be /mnt/cache on the templates.  /mnt/user works just fine.

 

Best guess on the broken images is that plex is referencing /mnt/user (you can verify by editing the container) and unRaid is grabbing the version of the file(s) on the array and missing the symlinks because only half the share got copied.

 

Back to what I originally suggested.  Delete /mnt/user0/apps and restore again or even better, delete /mnt/user/apps and restore again.  Either way you'll be back to where you where a couple of days ago.

 

awesome, I think I will give that a whirl in a day or so... because when sh*t hits the fan, it hits the fan and of course I had a drive die at the same time, so right now I am preclearing and adding a replacement in the mix, so I definitely dont want to do anything right now.

And if you have to deal with WAF then that consideration trumps all others.

 

*But* if you're not using the dated backup feature of CA, then I'd disable the backup schedule just so you don't wind up trashing the backup set from a couple of days ago until you get around to restoring again.

[wickedathletes]

And if you have to deal with WAF then that consideration trumps all others.

 

*But* if you're not using the dated backup feature of CA, then I'd disable the backup schedule just so you don't wind up trashing the backup set from a couple of days ago until you get around to restoring again.

 

Worse than WAF, I have my parents breathing down my neck because they are 3 episodes left of breaking bad and my server has been down 2 days. damn you remote access plex. hahaha