summaryrefslogtreecommitdiffstats
path: root/wiki/src/blueprint/lizard_failover.mdwn
blob: d892e680a9efe19559c35f9f2e5e95adf7e95d65 (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
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
This is about [[!tails_ticket 6185]] and friends.

[[!toc levels=2]]

# Scope

The idea is to avoid blocking the release process and normal Tails
usage if lizard goes down. It's *not* meant to replace lizard for more
than a few weeks.

# Services

These are the services, among [[those we
run|contribute/working_together/roles/sysadmins]], that we want
a failover for. This is inspired by the [[priority list for service
monitoring|blueprint/monitor_servers#services]].

- APT repository (our own one + the freezable Debian one)
- bittorrent
- Gitolite + git-annex for our Tor Browser archive
- rsync server
- WhisperBack relay

<a id="system-specs"></a>

# System requirements

This is about [[!tails_ticket 10243]].

The goal is to specify what we'll likely need until end of 2018.
We also need datacenter-class bandwidth.

- APT repository:
  * CPU: reprepro can already be hungry
  * RAM: currently low, but might grow for the same reason as CPU usage
  * storage: 1.8T
- bittorrent:
  * CPU: low
  * RAM: low
  * storage: 3 GB
- Tor Browser archive:
  * CPU: low
  * RAM: low
  * storage: 200G
- rsync server:
  * CPU: low
  * RAM: low
  * storage: 5 GB (excluding `obsolete`; 20 GB including it)
- WhisperBack relay:
  * CPU: low
  * RAM: low
  * storage: low

**Total**:

  * CPU: currently low but must be easily upgradable.
  * RAM: 2 GB should be enough for now, but must be easily upgradable
  * storage: 2T (preferabily RAID) to cover up until end 2018.