I'm using the default Amazon Linux AMI EBS snapshot, which is about 8 GB, but only handles half-a-million inodes (as per df -i) leading node_modules to rapidly exhaust disk space ("no disk space left on device" even when it still has 2 or 3 gigs "free"). This happens after 6-10 deploys.
Problem is ext4, I think. Easiest might be to add a separate XFS volume and use that. (EBS is cheap, currently about $0.10/GB-month, so I could add a 10GB volume for a buck a month.)
In any case, want a daily cron to clean up old deploys (and node_modules caches).
[x] Auto-cleanup old deploy directories and node_modules caches.
I'm using the default Amazon Linux AMI EBS snapshot, which is about 8 GB, but only handles half-a-million inodes (as per
df -i) leadingnode_modulesto rapidly exhaust disk space ("no disk space left on device" even when it still has 2 or 3 gigs "free"). This happens after 6-10 deploys.Problem is ext4, I think. Easiest might be to add a separate XFS volume and use that. (EBS is cheap, currently about $0.10/GB-month, so I could add a 10GB volume for a buck a month.)
In any case, want a daily cron to clean up old deploys (and
node_modulescaches).node_modulescaches.