ZFS snapshot automation tool

• cd /opt
• git clone https://github.com/makhomed/autosnap.git autosnap

Also you need to install python3:

# yum install python3

• cd /opt/autosnap
• git pull

• vim /opt/autosnap/autosnap.conf
• write to config something like this:

interval hourly 24
interval daily  15
interval weekly 8

exclude **-dump
exclude **-swap

ignore tank/backup/**
ignore tank/mirror/**

Configuration file allow comments, from symbol # to end of line.

Configuration file has only four directives: interval, exclude, include and ignore.

Syntax of interval directive: interval <name> <count>. <name> is name of interval, must be unique. <count> is count of snapshots to save for interval <name>.

Interval name can be frequent, hourly, daily, weekly, monthly, yearly or something else. Each interval name should be configured via cron for running.

Syntax of include, exclude and ignore directives are the same: exclude <pattern> or include <pattern> or ignore <pattern>.

By default all datasets are included. But you can exclude some datasets by name or by pattern. Pattern is rsync-like, ? means any one symbol, * means any symbols except / symbol, ** means any symbols.

First match win, and if it was directive exclude - dataset will be excluded, if it was directive include - dataset will be included.

exclude, include and ignore directives allowed only on global level.

ignore command works as exclude command, and also ignore datasets will be ignored in the autosnap -x command output.

interval directive allowed to be configured for each dataset separately. For example:

interval hourly 24
interval daily  15
interval weekly 8

[tank/kvm-stage-elastic]

interval hourly 24
interval daily  7
interval weekly 4

[tank/kvm-stage-mysqld]

interval hourly 24
interval daily  7
interval weekly 4

• vim /etc/cron.d/autosnap
• write to cron file something like this:

0 * * * * root /opt/autosnap/autosnap -x hourly
0 0 * * * root /opt/autosnap/autosnap -x daily
0 0 * * 7 root /opt/autosnap/autosnap -x weekly

At start autosnap read config from /opt/autosnap/autosnap.conf.

One and only one command must be specified in command line. This command must be the name of interval from configuration file.

During execution, autosnap will create one new snapshot for each included dataset and delete all oldest snapshots exceeding the allowed snapshots count for given interval.

If autosnap started without any parameters - then help printed and program will exit.

If autosnap started with -x parameter - it will list all snapshots, not managed by autosnap. The -x parameter can be combined with interval name.

If autosnap started with -d parameter - it will delete all expired snapshots, managed by autosnap. The -d parameter can not be combined with interval name.