xref: /freebsd/sbin/hastd/hast.conf.5 (revision 2e3f49888ec8851bafb22011533217487764fdb0)
1.\" Copyright (c) 2010 The FreeBSD Foundation
2.\" Copyright (c) 2010-2012 Pawel Jakub Dawidek <pawel@dawidek.net>
3.\" All rights reserved.
4.\"
5.\" This documentation was written by Pawel Jakub Dawidek under sponsorship from
6.\" the FreeBSD Foundation.
7.\"
8.\" Redistribution and use in source and binary forms, with or without
9.\" modification, are permitted provided that the following conditions
10.\" are met:
11.\" 1. Redistributions of source code must retain the above copyright
12.\"    notice, this list of conditions and the following disclaimer.
13.\" 2. Redistributions in binary form must reproduce the above copyright
14.\"    notice, this list of conditions and the following disclaimer in the
15.\"    documentation and/or other materials provided with the distribution.
16.\"
17.\" THIS SOFTWARE IS PROVIDED BY THE AUTHORS AND CONTRIBUTORS ``AS IS'' AND
18.\" ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
19.\" IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
20.\" ARE DISCLAIMED.  IN NO EVENT SHALL THE AUTHORS OR CONTRIBUTORS BE LIABLE
21.\" FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
22.\" DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
23.\" OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
24.\" HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
25.\" LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
26.\" OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
27.\" SUCH DAMAGE.
28.\"
29.Dd January 25, 2012
30.Dt HAST.CONF 5
31.Os
32.Sh NAME
33.Nm hast.conf
34.Nd configuration file for the
35.Xr hastd 8
36daemon and the
37.Xr hastctl 8
38utility
39.Sh DESCRIPTION
40The
41.Nm
42file is used by both
43.Xr hastd 8
44daemon
45and
46.Xr hastctl 8
47control utility.
48Configuration file is designed in a way that exactly the same file can be
49(and should be) used on both HAST nodes.
50Every line starting with # is treated as comment and ignored.
51.Sh CONFIGURATION FILE SYNTAX
52General syntax of the
53.Nm
54file is following:
55.Bd -literal -offset indent
56# Global section
57control <addr>
58listen <addr>
59replication <mode>
60checksum <algorithm>
61compression <algorithm>
62timeout <seconds>
63exec <path>
64metaflush on | off
65pidfile <path>
66
67on <node> {
68	# Node section
69        control <addr>
70        listen <addr>
71        pidfile <path>
72}
73
74on <node> {
75	# Node section
76        control <addr>
77        listen <addr>
78        pidfile <path>
79}
80
81resource <name> {
82	# Resource section
83	replication <mode>
84	checksum <algorithm>
85	compression <algorithm>
86	name <name>
87	local <path>
88	timeout <seconds>
89	exec <path>
90	metaflush on | off
91
92	on <node> {
93		# Resource-node section
94		name <name>
95		# Required
96		local <path>
97		metaflush on | off
98		# Required
99		remote <addr>
100		source <addr>
101	}
102	on <node> {
103		# Resource-node section
104		name <name>
105		# Required
106		local <path>
107		metaflush on | off
108		# Required
109		remote <addr>
110		source <addr>
111	}
112}
113.Ed
114.Pp
115Most of the various available configuration parameters are optional.
116If parameter is not defined in the particular section, it will be
117inherited from the parent section.
118For example, if the
119.Ic listen
120parameter is not defined in the node section, it will be inherited from
121the global section.
122In case the global section does not define the
123.Ic listen
124parameter at all, the default value will be used.
125.Sh CONFIGURATION FILE DESCRIPTION
126The
127.Aq node
128argument can be replaced either by a full hostname as obtained by
129.Xr gethostname 3 ,
130only first part of the hostname, by node's UUID as found in the
131.Va kern.hostuuid
132.Xr sysctl 8
133variable
134or by node's hostid as found in the
135.Va kern.hostid
136.Xr sysctl 8
137variable.
138.Pp
139The following statements are available:
140.Bl -tag -width ".Ic xxxx"
141.It Ic control Aq addr
142.Pp
143Address for communication with
144.Xr hastctl 8 .
145Each of the following examples defines the same control address:
146.Bd -literal -offset indent
147uds:///var/run/hastctl
148unix:///var/run/hastctl
149/var/run/hastctl
150.Ed
151.Pp
152The default value is
153.Pa uds:///var/run/hastctl .
154.It Ic pidfile Aq path
155.Pp
156File in which to store the process ID of the main
157.Xr hastd 8
158process.
159.Pp
160The default value is
161.Pa /var/run/hastd.pid .
162.It Ic listen Aq addr
163.Pp
164Address to listen on in form of:
165.Bd -literal -offset indent
166protocol://protocol-specific-address
167.Ed
168.Pp
169Each of the following examples defines the same listen address:
170.Bd -literal -offset indent
1710.0.0.0
1720.0.0.0:8457
173tcp://0.0.0.0
174tcp://0.0.0.0:8457
175tcp4://0.0.0.0
176tcp4://0.0.0.0:8457
177.Ed
178.Pp
179Multiple listen addresses can be specified.
180By default
181.Nm hastd
182listens on
183.Pa tcp4://0.0.0.0:8457
184and
185.Pa tcp6://[::]:8457
186if kernel supports IPv4 and IPv6 respectively.
187.It Ic replication Aq mode
188.Pp
189Replication mode should be one of the following:
190.Bl -tag -width ".Ic xxxx"
191.It Ic memsync
192.Pp
193Report the write operation as completed when local write completes and
194when the remote node acknowledges the data receipt, but before it
195actually stores the data.
196The data on remote node will be stored directly after sending
197acknowledgement.
198This mode is intended to reduce latency, but still provides a very good
199reliability.
200The only situation where some small amount of data could be lost is when
201the data is stored on primary node and sent to the secondary.
202Secondary node then acknowledges data receipt and primary reports
203success to an application.
204However, it may happen that the secondary goes down before the received
205data is really stored locally.
206Before secondary node returns, primary node dies entirely.
207When the secondary node comes back to life it becomes the new primary.
208Unfortunately some small amount of data which was confirmed to be stored
209to the application was lost.
210The risk of such a situation is very small.
211The
212.Ic memsync
213replication mode is the default.
214.It Ic fullsync
215.Pp
216Mark the write operation as completed when local as well as remote
217write completes.
218This is the safest and the slowest replication mode.
219.It Ic async
220.Pp
221The write operation is reported as complete right after the local write
222completes.
223This is the fastest and the most dangerous replication mode.
224This mode should be used when replicating to a distant node where
225latency is too high for other modes.
226.El
227.It Ic checksum Aq algorithm
228.Pp
229Checksum algorithm should be one of the following:
230.Bl -tag -width ".Ic sha256"
231.It Ic none
232No checksum will be calculated for the data being send over the network.
233This is the default setting.
234.It Ic crc32
235CRC32 checksum will be calculated.
236.It Ic sha256
237SHA256 checksum will be calculated.
238.El
239.It Ic compression Aq algorithm
240.Pp
241Compression algorithm should be one of the following:
242.Bl -tag -width ".Ic none"
243.It Ic none
244Data send over the network will not be compressed.
245.It Ic hole
246Only blocks that contain all zeros will be compressed.
247This is very useful for initial synchronization where potentially many blocks
248are still all zeros.
249There should be no measurable performance overhead when this algorithm is being
250used.
251This is the default setting.
252.It Ic lzf
253The LZF algorithm by Marc Alexander Lehmann will be used to compress the data
254send over the network.
255LZF is very fast, general purpose compression algorithm.
256.El
257.It Ic timeout Aq seconds
258.Pp
259Connection timeout in seconds.
260The default value is
261.Va 20 .
262.It Ic exec Aq path
263.Pp
264Execute the given program on various HAST events.
265Below is the list of currently implemented events and arguments the given
266program is executed with:
267.Bl -tag -width ".Ic xxxx"
268.It Ic "<path> role <resource> <oldrole> <newrole>"
269.Pp
270Executed on both primary and secondary nodes when resource role is changed.
271.It Ic "<path> connect <resource>"
272.Pp
273Executed on both primary and secondary nodes when connection for the given
274resource between the nodes is established.
275.It Ic "<path> disconnect <resource>"
276.Pp
277Executed on both primary and secondary nodes when connection for the given
278resource between the nodes is lost.
279.It Ic "<path> syncstart <resource>"
280.Pp
281Executed on primary node when synchronization process of secondary node is
282started.
283.It Ic "<path> syncdone <resource>"
284.Pp
285Executed on primary node when synchronization process of secondary node is
286completed successfully.
287.It Ic "<path> syncintr <resource>"
288.Pp
289Executed on primary node when synchronization process of secondary node is
290interrupted, most likely due to secondary node outage or connection failure
291between the nodes.
292.It Ic "<path> split-brain <resource>"
293.Pp
294Executed on both primary and secondary nodes when split-brain condition is
295detected.
296.El
297.Pp
298The
299.Aq path
300argument should contain full path to executable program.
301If the given program exits with code different than
302.Va 0 ,
303.Nm hastd
304will log it as an error.
305.Pp
306The
307.Aq resource
308argument is resource name from the configuration file.
309.Pp
310The
311.Aq oldrole
312argument is previous resource role (before the change).
313It can be one of:
314.Ar init ,
315.Ar secondary ,
316.Ar primary .
317.Pp
318The
319.Aq newrole
320argument is current resource role (after the change).
321It can be one of:
322.Ar init ,
323.Ar secondary ,
324.Ar primary .
325.It Ic metaflush on | off
326.Pp
327When set to
328.Va on ,
329flush write cache of the local provider after every metadata (activemap) update.
330Flushing write cache ensures that provider will not reorder writes and that
331metadata will be properly updated before real data is stored.
332If the local provider does not support flushing write cache (it returns
333.Er EOPNOTSUPP
334on the
335.Cm BIO_FLUSH
336request),
337.Nm hastd
338will disable
339.Ic metaflush
340automatically.
341The default value is
342.Va on .
343.It Ic name Aq name
344.Pp
345GEOM provider name that will appear as
346.Pa /dev/hast/<name> .
347If name is not defined, resource name will be used as provider name.
348.It Ic local Aq path
349.Pp
350Path to the local component which will be used as backend provider for
351the resource.
352This can be either GEOM provider or regular file.
353.It Ic remote Aq addr
354.Pp
355Address of the remote
356.Nm hastd
357daemon.
358Format is the same as for the
359.Ic listen
360statement.
361When operating as a primary node this address will be used to connect to
362the secondary node.
363When operating as a secondary node only connections from this address
364will be accepted.
365.Pp
366A special value of
367.Va none
368can be used when the remote address is not yet known (eg. the other node is not
369set up yet).
370.It Ic source Aq addr
371.Pp
372Local address to bind to before connecting to the remote
373.Nm hastd
374daemon.
375Format is the same as for the
376.Ic listen
377statement.
378.El
379.Sh FILES
380.Bl -tag -width ".Pa /var/run/hastctl" -compact
381.It Pa /etc/hast.conf
382The default
383.Xr hastctl 8
384and
385.Xr hastd 8
386configuration file.
387.It Pa /var/run/hastctl
388Control socket used by the
389.Xr hastctl 8
390control utility to communicate with the
391.Xr hastd 8
392daemon.
393.El
394.Sh EXAMPLES
395The example configuration file can look as follows:
396.Bd -literal -offset indent
397listen tcp://0.0.0.0
398
399on hasta {
400	listen tcp://2001:db8::1/64
401}
402on hastb {
403	listen tcp://2001:db8::2/64
404}
405
406resource shared {
407	local /dev/da0
408
409	on hasta {
410		remote tcp://10.0.0.2
411	}
412	on hastb {
413		remote tcp://10.0.0.1
414	}
415}
416resource tank {
417	on hasta {
418		local /dev/mirror/tanka
419		source tcp://10.0.0.1
420		remote tcp://10.0.0.2
421	}
422	on hastb {
423		local /dev/mirror/tankb
424		source tcp://10.0.0.2
425		remote tcp://10.0.0.1
426	}
427}
428.Ed
429.Sh SEE ALSO
430.Xr gethostname 3 ,
431.Xr geom 4 ,
432.Xr hastctl 8 ,
433.Xr hastd 8
434.Sh AUTHORS
435The
436.Nm
437was written by
438.An Pawel Jakub Dawidek Aq Mt pjd@FreeBSD.org
439under sponsorship of the FreeBSD Foundation.
440