Hi there,
I’m setting up a backup server running Amanda 3.3.6 on Ubuntu 16.04. My main
idea is to use the amzfs-sendrecv plugin. The dump part works, but I have a
couple of questions about the restore procedures. If anyone could give me some
advice that would be great.
I’m still having a hard time in figuring out the interplay between ZFS
snapshots and Amanda. As far as I can tell, Amanda create a temporary ZFS
snapshot, and send the delta between that and the previous one the backup. Then
it proceeds to swap / rename the snapshot for the next run. When a full backup
is needed a it sends the snapshots without the `-i` flag. Assuming my
understanding is correct, I still don’t know how to:
- execute a full restore a ZFS dataset, i.e. the dataset is gone. In
particular, I’d like to use amfetchdump so that I can pipe the stream back with
`zfs recv`
- execute a partial restore, i.e. a point in time recovery with an existing ZFS
dataset. Now, I realize that you could solve this particular use case just with
ZFS, assuming you have a usable ZFS snapshot, but since I’m just getting
started with Amanda, I’d like to know how to solve it with this tool, as well.
The configuration shown below is what I come up with by looking at the wiki
pages. I’d be really grateful if someone could give me feedback on this. Does
anybody see anything wrong or that can be improved?
My conf:
> infofile "/var/lib/amanda/state/curinfo"
> indexdir "/var/lib/amanda/state/index"
> dumpuser "backup"
> mailto “<…>"
>
> define changer s3 {
> tapedev
> "chg-multi:s3:<…>/slot-bb2e601e63b3408bb6a865e442a28366-{1,2,3,4,5,6,7,8,9,10,11,12,13,14,15}"
> changerfile "/var/lib/amanda/state/s3-statefile"
> device-property "NB_THREADS_BACKUP" "8"
> device-property "NB_THREADS_RECOVERY" "8"
> device_property "S3_ACCESS_KEY” “<…>"
> device_property "S3_SECRET_KEY” “<…>"
> device_property "S3_BUCKET_LOCATION" "eu-west-1"
> device_property "S3_SSL" "YES"
> device_property "BLOCK_SIZE" "10 megabytes"
> }
>
> tpchanger "s3"
> tapetype S3
>
> define tapetype S3 {
> comment "S3 Bucket"
> length 10240 gigabytes
> }
>
> org "zfs-dataset"
> logdir "/var/log/zfs-dataset"
>
> define application-tool zfs-dataset-app {
> comment "amzfs-sendrecv"
> plugin "amzfs-sendrecv"
> property "ZFS-PATH" "/sbin/zfs"
> property "PFEXEC-PATH" "/usr/bin/sudo"
> property "PFEXEC" "YES"
> }
>
> define dumptype zfs-dataset-dump {
> program "APPLICATION"
> application "zfs-dataset-app"
> auth "ssh"
> ssh_keys "/var/backups/.ssh/id_rsa”
> }
--
Giorgio Valoti
