Ram Shoots to 100% then Kills sync

I’m running a VM for rclone and have 6cores/8gb ram on 10gb HDD provisioned. I run the following command:

rclone copy /mnt/foldera crypt:folderb --transfers=10 --drive-chunk-size=1G --drive-upload-cutoff=1G --drive-use-trash --stats=15s -vv

After running for a bit I get an error saying Killed. Nothing more in the log.

Here is the last transfer log:

Transferred: 6.499 GBytes (69.570 MBytes/s)
Errors: 0
Checks: 131
Transferred: 1
Elapsed time: 1m35.6s

And here is what it says right afterwards

2017/04/08 20:49:27 INFO : fileaaaa.gz: Copied (new)

Does anyone know why it dies. I guess I’m not understanding the drive chunk / upload cutoff? If I set it lower to 512M or 256M it doesn’t occur.

I’ve had this before. Reduce your transfers to 5 or use the default (4).

Do you know what the issue is? Why does it happen. Is it related to transfer numbers or the drive-chunk / upload-cutoff settings?

This is the size of chunks buffered in memory before uploading. So setting this to 1G will use --transfers * 1G of memory at least!

Set it much smaller would be my advice.[quote=“Omitted, post:1, topic:1672”]

What you are saying here is don’t used chunked upload for files < 1 GB. I wouldn’t bother changing this particularly.

Thanks, that explains it very clearly!