zfs send/receive aborts half way

[seldo]

I have a filesystem that comes from pool that was set up on another system (NAS4Free 11).
I'd like to replicate that file system to my current system.
The latest snapshot was done manually on FreeNAS 11.3

I do something as simple as:

Code:

zfs send BACKUP_POOL/migration/MAIN_POOL/photovideo/masters@migration_ready | zfs receive ARCHIVE_POOL/photos/masters

But the transfer doesn't complete.

From zfs list -r BACKUP_POOL: I have 1.48 TB to transfer

Code:

BACKUP_POOL/migration/MAIN_POOL/photovideo/originaux       1.48T   983G  1.45T  /mnt/BACKUP_POOL/migration/MAIN_POOL/photovideo/originaux

Here's what I see in the reporting page: it ran for about 2h20, at about 100 MiB/s
Calculating, that's 60*60*2.33 ≈ 8400 seconds * 100 = 840 MiB transferred.
So that's about 37% of my data.
I couldn't find anything in the logs in /var/log

Since the transfer aborts, the filesystem that is created at the beginning of the transfer is destroyed when the transfer doesn't finish.

Here's my memory report. I did a reboot before the transfer. I only have 16 GB (my system doesn't support more).
I currently don't have swap (the drives where I will enable it are not yet erased/recreated).

Is that only caused by a lack of memory?
Is it expected that I don't find any trace of the send/receive failure?
How can I log my future send/receive?

That's the only large filesystem I have to import.
I will never transfer such an amount in the future.

 

[seldo]

I've added a drive in USB, put 99 GiB of swap and rerun the same command.
Let's see if that works.

p.s: 99 GiB is the max value for swap, though it only shows red when above, but doesn't tell you what is the max value.