Register for the iXsystems Community to get an ad-free experience and exclusive discounts in our eBay Store.

SOLVED Need help with copying 30tb file on truenas

Western Digital Drives - The Preferred Drives of FreeNAS and TrueNAS CORE

Truenasser

Cadet
Joined
Jun 29, 2022
Messages
8
Hi there!
I am trying to use a zfs raidz2 pool as repository for Veeam, i have a backup of a physical server to do.it has already failed (after 5 days of data transfer) because the single file size exceeds the 15tb.
I've tried to have a look around but haven't find a solution about the issue.
Anyone who could point me to the right direction?
Thanks!!
 

sretalla

Hall of Famer
Joined
Jan 1, 2016
Messages
6,811

Truenasser

Cadet
Joined
Jun 29, 2022
Messages
8
Hi Sretalla thanks for your reply.
The error message is from Veeam, so the copying tool, not from Truenas.
 

sretalla

Hall of Famer
Joined
Jan 1, 2016
Messages
6,811
OK, I would start by looking into limitations from the software then... ZFS (and TrueNAS) allow files to be as large as 16 exbibytes (much larger than the limit you're experiencing there).

I'm not sure if SAMBA (I guess you're using an SMB share) has any limits like that, but I would expect not (unless you set one).
 

Truenasser

Cadet
Joined
Jun 29, 2022
Messages
8
Could it be because of the LZ4 compression on the Truenas?? I dont think i have set any filesize restrictions either on Veeam or Truenas. On Truenas just added pool/Raidz2/ZFS etc using default settings. Record size is 128K.
Thanks!
 

sretalla

Hall of Famer
Joined
Jan 1, 2016
Messages
6,811
As far as I can tell, lz4 compression is streamed, so there should be no limit coming from that.

You could try changing the dataset properties either to ZSTD or OFF and see if that changes anything... I would put that in the category of 10% chance of being any different.
 

sretalla

Hall of Famer
Joined
Jan 1, 2016
Messages
6,811
I would be checking these out if I were you:

How are you connecting from Veeam to the TrueNAS server? (iSCSI, SMB, NFS?)
 

sretalla

Hall of Famer
Joined
Jan 1, 2016
Messages
6,811

Truenasser

Cadet
Joined
Jun 29, 2022
Messages
8
could it be because of the share parameters set to "default" - maybe i have to pick some other paramenter?
couple screenshot attached.

Thanks again for your time, much appreciated!
 

Attachments

  • veeam error.png
    veeam error.png
    102.6 KB · Views: 17
  • share parameters.png
    share parameters.png
    60.1 KB · Views: 21

Truenasser

Cadet
Joined
Jun 29, 2022
Messages
8
ow i didnt see your previous replies with links to a possible solution, i will look them right now thanks again!
 

Truenasser

Cadet
Joined
Jun 29, 2022
Messages
8
I would be checking these out if I were you:

How are you connecting from Veeam to the TrueNAS server? (iSCSI, SMB, NFS?)
using SMB Sretalla.
I had a look but didnt find a hint on these links you've provided (or maybe i'm just blind).
Asked on Veeam forums and they said:
"Many of similar storages as TrueNAS have some limitations independent from the used filesystem maximum file sizes. So please reach out to the TrueNAS vendor"
 

sretalla

Hall of Famer
Joined
Jan 1, 2016
Messages
6,811
"Many of similar storages as TrueNAS have some limitations independent from the used filesystem maximum file sizes. So please reach out to the TrueNAS vendor"
@anodos any ideas on something from SAMBA that might be limiting file site to 15TB?
 

anodos

Sambassador
iXsystems
Joined
Mar 6, 2014
Messages
8,547
@anodos any ideas on something from SAMBA that might be limiting file site to 15TB?
Not that I'm aware of. SMB2 write requests are basically a file handle, offset, bytes count, and buffer. On the backend it just gets converted into a pwrite() or aio_write() request (or on SCALE an io_uring request). If you think there's a samba issue it is generally prominently visible in /var/log/samba4/log.smbd. Look there. Otherwise look in other logs for issues. With a 30 TiB single file transfer unless application has good error handling you run into risk that a network disruption (say mucking around with stuff in jails, plugins, etc) can disrupt it. Typically a durable reconnect should be possible, but things can happen (once again check your logs).
 
Top