RSYNC stalled?

[electricd7]

Hello,

I have a couple of RSYNC jobs setup that send data from my FreeNAS to another server. They seemed to be working great, but now appear to be stalled? I have --verbose logging turned on so when I look at the log I see lots of lines like the following:

2015/01/07 15:29:14 [7171] <f+++++++++ video/folder1/File/file.avi

But when I look at the network graph in the GUI there is barely any traffic. It used to be pegged around 100M. Also when I look at the receiving end of the copy, the folder structure is there but I don't see any files. There appear to be files in the .~tmp~ directory on the receiving end, but not where they should be. What is happening here?

 

[Joerg G]

Hi,
I was wondering as well the other day about these .~tmp~ directories containing the files I wanted to sync.
However after successful run of the complete rsync batch the files have been moved to the correct folder at the end.
So I guess this is due to the generated rsync options from freenas, I read that you can also configure it to directly write to the corresponding file.

As I said I was wondering as well and have therefore setup a test sync of a much smaller folder in order to check how it will look like in the end.

Cheers,
Joerg

 

[electricd7]

Ok thanks for the explanation. My stalled problem seems to persist though. My issue is that I have a failing FreeNAS so I am trying to backup all the data on it via Rsync to another server. Everything works great until my FreeNAS box reboots (which it does at least 2-3 times per day). Once that happens, the RSYNC task is scheduled to start every 15 minutes. The next time it kicks off after a reboot, I never see data start transferring again. Its like it tries to create a list of what to transfer and then never starts again. How do I force it to pick up where it left off?

 

[Joerg G]

Why not simply remove the disks and attach them to the other machine and import ZFS volume if the old machine is not working any longer...?

 

[electricd7]

sorry, I should have been more clear. Its the disks which are causing the reboot. I plan to backup all the data off the pool, and then replace all the disks with new ones and then move the data back onto the new pool.

 

[Joerg G]

OK, got it now...

Have you triggered rsync run from freenas web GUI or from ssh shell?
Maybe you have more control via shell and can use options like --inplace or others to skip using temp directories...

 

[electricd7]

I triggered it from within the GUI. How would I do it from within the shell to accomplish what I want?

 

[electricd7]

I think what happens is it transfers a couple hundred gigabytes which is thousands of files, then when it reboots it has to scan the far-end directory to see what is there and what is not so that it can start transferring again, but it takes so long to enumerate those files that it never starts copying again before the machine reboots again.

 

[Joerg G]

You can log in to freenas via ssh if you enable ssh service.
There you can trigger rsync command manually and pass desired options.
Maybe also try to rsync smaller directories in chunks to get around this issue of completely rebuilding the directory/file tree every time.