summaryrefslogtreecommitdiff
path: root/TODO.md
blob: a3607fec32525ebb8bac52e8e36683c97e943cd0 (plain)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
TODO list for byteback
======================

* byteback-restore program

* byteback-prune does not cope with the case where the mount point is not
    writeable (should print error explaining that)

* retrieve scripts for checking status etc. from Bytemark managed hosts,
    integrate into main project.

* byteback-setup-client should support use of own accounts on server (i.e.
    not forcing you to ssh to byteback@server, change password etc.)

* byteback directory should be at the end of the load path, not start

* change default retry parameters to cover 24 hour window

* "backup could not be marked as complete" message unhelpful on client system - why?
    (should retry when SSH error == 255, otherwise give up)

* how do we stop backups from crashing server's kernel?
  * locking up completely (e.g. "touch newfile" never returns)
  * starting to run very very slowly until a btrfsck / remount
  * often nothing to do but "echo b > /proc/sysrq-trigger"

* give nilfs / zfs a go as alternatives?

* try to deal with https://btrfs.wiki.kernel.org/index.php/Problem_FAQ#I_get_.22No_space_left_on_device.22_errors.2C_but_df_says_I.27ve_got_lots_of_space ?

* out-of-date check should suspend judgment when backup is in progress

* add idea of progress on server side?

* report differences between backups (new, deleted, changed files)

* spotting a /var/lib/mysql directory and making a safe snapshot and re-copy
  of a MySQL data directory (using FLUSH TABLES WITH READ LOCK)

* (same for postgres using pg_start_backup() and pg_stop_backup())