Commit Graph

85 Commits

Author SHA1 Message Date
Girish Ramakrishnan
7bc5ae17cc Use unbound as nameserver
DO uses Google nameservers by default. This causes RBL queries to fail.

Can be tested with the following command:
$ host 2.0.0.127.zen.spamhaus.org
Host 2.0.0.127.zen.spamhaus.org not found: 3(NXDOMAIN)

With unbound we get:
$ host 2.0.0.127.zen.spamhaus.org
2.0.0.127.zen.spamhaus.org has address 127.0.0.2
2.0.0.127.zen.spamhaus.org has address 127.0.0.10
2.0.0.127.zen.spamhaus.org has address 127.0.0.4

Also, we do not use dnsmasq because it is not a recursive resolver. It will
always forward and this defaults to the value in /etc/network/interfaces
(which is Google DNS on DO!).
2016-08-11 10:32:54 -07:00
Johannes Zellner
65994f307f Make infra_version.js option and fix base image on DO 2016-08-10 12:45:23 +02:00
Johannes Zellner
c6c36a4f3c Also make box-setup.service depend on cloud-init for ec2 2016-06-22 10:21:56 +02:00
Johannes Zellner
2a3640032f Remove obsolete SELFHOSTED env 2016-06-22 10:21:56 +02:00
Girish Ramakrishnan
d51d2e5131 start addons and apps in the cloudron network
also remove getLinkSync, since we don't use linking anymore
2016-06-17 09:18:10 -05:00
Johannes Zellner
984a191e4c Use the variable correctly 2016-06-09 14:24:53 +02:00
Girish Ramakrishnan
aff9ff47bc use infra_version.js in baseimage script 2016-05-24 13:23:38 -07:00
Girish Ramakrishnan
83a252bd20 there is only mail container now 2016-05-15 21:15:53 -07:00
Girish Ramakrishnan
560f460a32 rename to sendmail 2016-05-13 20:48:31 -07:00
Girish Ramakrishnan
5322ed054d reserve 4190 for sieve 2016-05-13 18:48:05 -07:00
Girish Ramakrishnan
646ba096c3 start recvmail addon in setup_infra 2016-05-11 08:55:51 -07:00
Girish Ramakrishnan
5afff5eecc open 25 (inbound smtp) and 587 (inbound submission) 2016-05-11 08:48:50 -07:00
Girish Ramakrishnan
84206738e1 open port 993 (imap) 2016-05-11 08:47:59 -07:00
Girish Ramakrishnan
8837cc5a3c update nginx version 2016-04-29 19:38:06 -07:00
Girish Ramakrishnan
f2545e3def bump systemd to 229 2016-04-29 19:18:31 -07:00
Girish Ramakrishnan
7a72bf3f78 select mysql 5.7 2016-04-29 19:12:20 -07:00
Girish Ramakrishnan
cfb56d7eee install aws-cli tool (for backups) 2016-04-08 23:58:07 -07:00
Girish Ramakrishnan
f44c2707f0 install swaks in base image 2016-04-04 09:50:19 -07:00
girish@cloudron.io
61a52d8888 dist-upgrade to update more aggressively 2016-03-09 09:29:07 -08:00
Girish Ramakrishnan
83d05c99d3 mount manually instead of fstab because of race
I cannot figure how to make the box-setup.service run before the mounting
of a specific mount point. adding a dep on mount.target locks up the system.
2016-03-07 10:48:09 -08:00
Girish Ramakrishnan
b0acdfb908 use truncate instead of fallocate 2016-03-07 10:44:35 -08:00
Girish Ramakrishnan
b062dab65c mysql also uses the data partition 2016-03-07 10:38:59 -08:00
girish@cloudron.io
744c721000 use docker 1.10.2 (untested) 2016-03-01 10:13:44 -08:00
girish@cloudron.io
0500bae221 install aufs tools
https://github.com/docker/docker/issues/915
2016-03-01 10:13:04 -08:00
girish@cloudron.io
1419108a86 umount is for unmounting 2016-02-25 20:13:16 -08:00
girish@cloudron.io
ee9c8ba4eb fix dead comment 2016-01-25 20:15:54 -08:00
girish@cloudron.io
d56abd94a9 collectd uses the data lo partition that is resized by box-setup.sh 2016-01-22 15:06:43 -08:00
girish@cloudron.io
a246b3e90c box-setup needs to run after mounting to prevent race in script 2016-01-22 14:21:36 -08:00
girish@cloudron.io
e28e1b239f fix comment 2016-01-22 14:21:20 -08:00
girish@cloudron.io
a763b08c41 pin packages
fixes #558
2016-01-22 10:46:13 -08:00
Girish Ramakrishnan
bb88fa3620 Restart node processes if journald crashes
Note that we cannot simply ignore EPIPE in the node programs.
Doing so results in no logs anymore :-( This is supposedly
fixed in systemd 228.

Fixes #550
2016-01-21 22:13:19 -08:00
Girish Ramakrishnan
1e1249d8e0 Give journald more time to sync
Part of #550
2016-01-21 21:43:49 -08:00
girish@cloudron.io
bcb0e61bfc Kill child processes
On Unix, child processes are not killed when parent dies.

Each process is part of a process group (pgid). When pgid == pid,
it is the process group leader.

node creates child processes with the parent as the group leader
(detached = false).

You can send a signal to entire group using kill(-pgid), as in,
negative value in argument. Systemd can be made to do this by
setting the KillMode=control-group.

Unrelated: Process groups reside inside session groups. Each session
group has a controlling terminal. Only one process in the session
group has access to the terminal. Process group is basically like
a bash pipeline. A session group is the entire login session with only
one process having terminal access at a time.

Fixes #543
2016-01-21 17:44:17 -08:00
Johannes Zellner
f82e714b3c Remove executable flag for scripts not intended to be called directly 2016-01-13 10:58:20 +01:00
Johannes Zellner
9d45892603 Move baseimage relevant scripts to baseimage/ 2016-01-13 10:30:01 +01:00