Jump to content
pjneder

Mover not working, files stuck on cache drive

12 posts in this topic Last Reply

Recommended Posts

Still having a few glitches getting going with 4.5 Plus on my 5 drive system. I was running Basic with just 2 drives 1 parity, 1 data.

 

My setup now is:

1x SATA 1TB Parity HDD

3x SATA 1TB data HDDs

1x IDE/PATA 80GB cache HDD

 

I have a few user shares all set to non-cache (files, backup, etc) on disk3. I have 1 share (Media) setup as:

include: disk1,disk2

exclude: disk3

cache: yes

 

If I write something to the Media share it goes onto the cache drive. However, 'mover' is not transferring it automatically via cron and manually clicking "move now" still does not result in the files going from /mnt/cache to /mnt/disk(1,2).

 

Mover logging is enabled but I cannot figure out where the mover log lives.

 

Hints?

 

Thanks!

Share this post


Link to post

OK. Right after I posted this I used different search words and found some posts that indirectly helped me. I still don't understand exactly what happened but at least mover is working now.

 

I found that "logger" posts to /var/log/syslog. So now I know where to look.

 

I found that mover was throwing "No space left on device (28)" errors. I set the min free space to 0 and it started working. See below....

 

So, here is more about the Media share:

 

Allocation method: Fill-up

Min free space: (This was set to 10737418240, for 10GB) now is set to "0"

Split level:

Included disk(s): disk1,disk2

Excluded disk(s): disk3

Use cache: yes

Export (SMB): export read/write

Export (NFS):

 

 

So, it was my understanding that if I used "Fill-up" and set the size to 10GB it would fill the drive until there was only 10GB left and then start filling disk2. Did I do something wrong?

 

Thanks

Share this post


Link to post

Still having a few glitches getting going with 4.5 Plus on my 5 drive system. I was running Basic with just 2 drives 1 parity, 1 data.

 

My setup now is:

1x SATA 1TB Parity HDD

3x SATA 1TB data HDDs

1x IDE/PATA 80GB cache HDD

 

I have a few user shares all set to non-cache (files, backup, etc) on disk3. I have 1 share (Media) setup as:

include: disk1,disk2

exclude: disk3

cache: yes

 

If I write something to the Media share it goes onto the cache drive. However, 'mover' is not transferring it automatically via cron and manually clicking "move now" still does not result in the files going from /mnt/cache to /mnt/disk(1,2).

 

Mover logging is enabled but I cannot figure out where the mover log lives.

 

Hints?

 

Thanks!

It logs to the syslog.  ( /var/log/syslog )

 

Looking at the "mover" script itself (in /usr/local/sbin/mover ), it will not move anything if the cache disk is not enabled (/mnt/user0 directory must exist)

 

It will not run again if it is already running.

 

It will not move any file if it is currently open and still being written to.

 

It will not move anything in a folder name that starts with a leading period "."

Share this post


Link to post

So, it was my understanding that if I used "Fill-up" and set the size to 10GB it would fill the drive until there was only 10GB left and then start filling disk2. Did I do something wrong?

 

Thanks

You need to set the split level.  If you wish for it to automatically create the top level folder, set it to 1.  If you want it to create the top two levels of folders, set it to 2, etc....  If you want to create the folders yourself on disk2, set it to 0.

 

Once you set the split level you can put the "Min free space" back to 10Meg.

Share this post


Link to post

Added to the FAQ (link back only).  Feel free to edit or expand.

  Why are my files not moving off the Cache drive?

 

We probably should add that there have been several cases when files would not move because of insufficient space on a specific drive due to the current split-level and allocation settings.  The FAQ entry "Why is my User Share out of space when there is plenty left?" may be helpful.

Share this post


Link to post

This problem has come back. Due to a power outage my unRAID went down recently, but none of the settings had changed.

 

I have a User share for Media that includes disk1/disk2. Split level=2 with Fill-up method. Min space set to 10GB.

 

The syslog shows that it thinks there is no disk space left on the share, which is not true.

 

To get the files off the cache I've set the min space back to 0 and manually run mover and it works.

 

Any ideas?

 

Thanks,

Paul

Share this post


Link to post

This problem has come back. Due to a power outage my unRAID went down recently, but none of the settings had changed.

 

I have a User share for Media that includes disk1/disk2. Split level=2 with Fill-up method. Min space set to 10GB.

 

The syslog shows that it thinks there is no disk space left on the share, which is not true.

 

To get the files off the cache I've set the min space back to 0 and manually run mover and it works.

 

Any ideas?

 

Thanks,

Paul

Exactly what are you entering to specify 10GB min-space?  What exactly is in the field on the form?

What free space exists on your drives when you issue the command

df

 

Since you did not post the syslog, can't determine what you are seeing.  I personally do not use a cache drive, never have, probably never will.

 

My instinct is that the min-free space setting is involved, but then, as I said, there have been more complaints about the cache drive and space allocation ans split level than almost anything else.  I allocate my own space, as I need and write to disk shares directly and do not rely on unRAID to pick a disk for me.

 

Best to address this directly with Tom @ lime-technology.

Share this post


Link to post

Exactly what are you entering to specify 10GB min-space?

What exactly is in the field on the form?

 

As per my earlier post in this thread:

10737418240

 

What free space exists on your drives when you issue the command

df

 

This is the most recent, AFTER I just copied five ~6GBish movie files over to disk1 manually using /usr/local/sbin/mover

Floater login: root
Password:
Linux 2.6.31.6-unRAID.
root@Floater:~# df
Filesystem           1K-blocks      Used Available Use% Mounted on
/dev/sde1              1954036    371440   1582596  20% /boot
/dev/hda1             80415748   8136992  72278756  11% /mnt/cache
/dev/md2             976732736     32840 976699896   1% /mnt/disk2
/dev/md3             976732736 290855756 685876980  30% /mnt/disk3
/dev/md1             976732736 951248224  25484512  98% /mnt/disk1
shfs                 3010613956 1250273812 1760340144  42% /mnt/user
shfs                 2930198208 1242136820 1688061388  43% /mnt/user0
root@Floater:~#

 

Since you did not post the syslog, can't determine what you are seeing.  I personally do not use a cache drive, never have, probably never will.

 

My instinct is that the min-free space setting is involved, but then, as I said, there have been more complaints about the cache drive and space allocation ans split level than almost anything else.  I allocate my own space, as I need and write to disk shares directly and do not rely on unRAID to pick a disk for me.

 

Best to address this directly with Tom @ lime-technology.

 

I will consider sending him an e-mail directly as well. Yes, this has caused lots of problems and is non-intuitive despite the documentation.

 

Thanks,

Paul

Share this post


Link to post

Min free is in 1024byte blocks.  see here:

http://lime-technology.com/forum/index.php?topic=3760.msg32767;topicseen#msg32767

 

Your field value of

10737418240 = 10737418240 * 1024 bytes or

10,995,116,277,760 bytes.

 

10,995,116,277,760 bytes  / 1024 =  10,737,418,240 kilobytes

10,737,418,240  kilobytes / 1024 =        10,485,760 megabytes

10,485,760 megabytes / 1024 =            10,240 gigabytes

10,240 gigabytes / 1024 =                10 terabytes

 

It looks to me as if you've asked that at least 10TB be free on the target disk before the mover uses it.

Was that your intent?  ;)  None of your disks look to have that amount of free space.  ;D

 

I think if you adjust the min-free to something like 10485760 it might just start to work. (That is 10GB in 1024k blocks)

Share this post


Link to post

Oh SNAP!  :o :o :o

 

I failed completely to see the notation about 1024k blocks. I know how to do the math, but not know it was 1kB chunks put me off an order of magnitude. No wonder is was all dorked up!

 

I'll fix that setting and see if things behave better.

 

Thanks Joe!

 

Share this post


Link to post

Oh SNAP!  :o :o :o

 

I failed completely to see the notation about 1024k blocks. I know how to do the math, but not know it was 1kB chunks put me off an order of magnitude. No wonder is was all dorked up!

 

I'll fix that setting and see if things behave better.

 

Thanks Joe!

 

Now you can write to lime-tech and complain that the field is not labeled properly... not that the mover does not work.  (It really should give the units involved)

Share this post


Link to post

(It really should give the units involved)

+1 ! I also stumbled over this in the past - at least a flyover or help/description should be made available ...

Share this post


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.