I'm replacing my current SSD cache drive which had about 35GB of data. I turned off the caching for all shares, stopped dockers,VM's disabled, etc and started mover this morning about 8am. It's 5PM and it's still running with 25.1GB on the drive. I know Plex has thousands and thousands of small files, but this seems very excessive to me. Is this abnormal?

I'm not seeing errors anywhere and it's moving data, but something doesn't seem right to me.

 

unRAID server:

ASUS TS700-E7/RS8 Server

32GB ECC RAM (4 x 8GB ECC DDR3L 1600)

2 x Xeon E5-2690 SR0L0 8-Core 2.90 20MB CPU's

Cache drive is a Corsair Neutron XT 240GB SATA III MLC SSD

It's moving everything to Disk 1 which is a 3TB Seagate SATA 6Gb/s 64MB cache drive (ST3000DM001).

 

I'm not using a parity drive. I have a CineRAID USB 3 8TB RAID 0 array and a Seagate 8TB USB 3 Backup Plus Hub for backups.

Edited June 23, 20179 yr by skeil909

21 minutes ago, skeil909 said:

ST3000DM001

https://en.wikipedia.org/wiki/ST3000DM001

 

Probably about right for moving a plex appdata to an older spinner drive with questionable history. In your specific case, it may have been better to put the replacement cache drive in as an array drive temporarily since you aren't running a parity drive. What file system is on the current cache disk, and what fs is the array disk?

16 minutes ago, jonathanm said:

https://en.wikipedia.org/wiki/ST3000DM001

 

Probably about right for moving a plex appdata to an older spinner drive with questionable history. In your specific case, it may have been better to put the replacement cache drive in as an array drive temporarily since you aren't running a parity drive. What file system is on the current cache disk, and what fs is the array disk?

 

I left them on the defaults when I first setup the system. Cache disk is btrfs and array disk is xfs.

 

As for the ST3000DM001, I have 4 of them. Bought them brand new back in November of 2012 and they've never had a problem. They've been running in a CineRAID as a RAID 5 array and only ran a few hours a day, a few times a week. No SMART errors or any issues with them at all.

Edited June 23, 20179 yr by skeil909

Probably would have been faster to add the new cache drive to the cache pool, let it balance the RAID1 pool, then remove the old cache and convert back to single profile.

 

Now that the mover is in progress, I'd keep an eye on it and make sure you keep seeing progress. It would be a good idea to keep an eye on the syslog for errors as well.

3 minutes ago, jonathanm said:

....

 

 

Thanks. I'll try to remember that if I ever need to do this again.

Edited June 23, 20179 yr by skeil909

Once mover had gone from 35GB down to about 20GB it only took about an hour to move the last 20GB. That first 15GB took about 15 hours. 

 

 

28 minutes ago, skeil909 said:

Once mover had gone from 35GB down to about 20GB it only took about an hour to move the last 20GB. That first 15GB took about 15 hours. 

 

 

Mover on 6.3.x is very inefficient and very slow especially with small files due to the major overhead it imposes.  (and that time frame is roughly correct)   6.4 is supposedly much faster.  As an aside, the appdata backup plugin would have done this in ~2 hours

5 hours ago, Squid said:

Mover on 6.3.x is very inefficient and very slow especially with small files due to the major overhead it imposes.  (and that time frame is roughly correct)   6.4 is supposedly much faster.  As an aside, the appdata backup plugin would have done this in ~2 hours

 

Ahh, I didn't even think about the appdata backup plugin. I just discovered this fantastic software a couple weeks ago and am still learning the in's and out's. Thanks for the info!

 

I bumped up my ram from 32gb to 64gb quad channel today and added 2 x 500gb 850 evo's for the cache pool. I also went ahead and added a parity drive.

Everything is running very well and it's so much nicer than Hyper-V. I'm hooked and can't believe I had never ran across this software before, It's great!