Jarrod Igou's picture

Hi, folks.

I'm trying to restore / migrate a website to a medium.m3 EC2 instance, but I'm running into issues with running out of disk space when I try to restore.

The medium.m3 instance has 4GB of SSD-based instance storage. I've also attached a 30GB SSD-based EBS volume and set that as the boot volume.

My backup is a 12-ish GB backup, and when I trigger the backup from my EC2 instance it runs out of space at /tmp because it's using the instance-based 4GB volume and not the 30GB EBS volume.

Do you guys have any thoughts or recommendations to help me get around this?



Jeremy Davis's picture

Firstly if you mount your 30GB to /mnt (or perhaps it was /mnt/tmp) then AFAIK TKLBAM should notice that and use that instead of /tmp (see this thread and this thread - they're old but describe issues similar to yours and have posts by the core devs explaining how to work around, plus the changes made to TKLBAM to try to prevent issues like this).

The other option would be to do a 2 step process. Download the backup first (to where ever you have the 30GB mounted) and then apply it. Like this: [source: tklbam-restore man]

tklbam-restore --raw-download=/path/to/30gb
tklbam-restore /path/to/30gb
Jarrod Igou's picture

Thanks, Jeremy! I had seen those other messages, but the two-step process I hadn't found.

In the end, since I had nothing configured or implemented on the machine, I decided to start over. This time the system set the boot volume up more sanely (OK, I might have made a couple different choices when I deployed this one) and I've got the space I need on my root volume. A df -H showed me that indeed this second time I've got the disk space I need in /tmp.

Currently running the restore process, I should be good to go shortly.

Thank you again!


Add new comment