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
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
|
@ COPYRIGHT Command Prompt, Inc.
pitr_tools is a suite of utilities to make point in time easier. It uses a
push mechanism for updating archives.
Options:
-h, --help show this help message and exit
-C FILE, --config=FILE
the name of the archiver config file
-f, --flush Flush all remaining archives to slave
Dependencies:
key based ssh auth between the master and slave (both ways)
To enable, one must edit the postgresql.conf.
archive_mode = on # allows archiving to be done
archive_command = 'bin/archiver -C etc/archiver.ini -F %p'
-C takes the path to the config file
-F takes the path (or file) of the archive.
-f/--flush is used at the command line only.
--flush:
You use flush before passing -F999 to cmd_standby. This is done
to make sure that all logs that master has written are available to
the standby before a failover.
Note that the use of --flush is a one way trip. Once used, you must
bring your slave into production or you will have to issue a new base
backup.
--init
Used to create the appropriate archive/queue file directories
--push:
The push option is designed to work with the protocol: file option.
Essentially this allows the archiver to copy the archive files to the
l_archivedir location. You would then call the archiver independently
of PostgreSQL using a scheduler such as cron. When you call the archiver
with push it will rsync (using ssh) all files in l_archivedir to
r_archivedir.
The archiver.ini file contains a variety of options.
[DEFAULT]
; online or offline
state: online
You can change the archiver to offline live, without touching postgresql.
When taking the archiver offline this will cause an non zero exit status
to postgresql's archive command on each archive attempt. This is o.k.
because postgresql will just queue up the archives while the archiver is
offline. However this can cause you to accidently fill up your hard disk
so use with care.
; The base database directory
pgdata: /data1/pgsql/data
This is the PGDATA on the archiving machine.
; where to remotely copy archives
r_archivedir: /data2/pgsql/archive
The remote archive directory, e.g; the directory that will accept all
archive files to be restored by the standby server.
; where to locally copy archives
l_archivedir: /data2/pgsql/archive
The local archive directory. This is used when you want to archive all
files locally. For example, if you are using an NFS mount for the standby
server to get its files. It is also used as a queue for when archiving fails
to a particular slave
Must point to the actual binary of scp.
; where is rsync
rsync_bin: /usr/bin/rsync
Must point to the actual binary of rsync. We use rsync over ssh, so don't
bother with an rsync server.
; IP of slave
slaves: 192.168.3.1 or slaves: 192.168.3.1,192.168.4.1
This is the ip address of the standby server. A hostname/FQDN would work as well.
It must be a command separated list for multiple slaves.
; the user that will be using rsync
user: postgres
The unix shell user that is using rsync. This is the user that must have ssh
keys configured between the master and slave.
; if rsync can't connect in 10 seconds error
timeout: 10
The SSH timeout. If it takes > timeout to begin the transfer the archiver will
provide a non zero return and alert per notify_critical.
; command to process in ok
notify_ok: echo ok
Everything is o.k. so what do we do? This should be full path to a script.
For example, a nagios alert.
; command to process in warning
notify_warning: echo WARNING
The same as OK but for warnings.
; command to process in critical
notify_critical: echo CRITICAL
The same as OK but for CRITICAL. This is total failure conditions.
; if you want to debug on/off only
debug: off
This will provide you a lot of noisy info to wherever you push your
postgresql logging.
|
