]>
Commit | Line | Data |
---|---|---|
22653ac8 DM |
1 | [[chapter-ha-manager]] |
2 | ifdef::manvolnum[] | |
b2f242ab DM |
3 | ha-manager(1) |
4 | ============= | |
22653ac8 | 5 | include::attributes.txt[] |
5f09af76 DM |
6 | :pve-toplevel: |
7 | ||
22653ac8 DM |
8 | NAME |
9 | ---- | |
10 | ||
734404b4 | 11 | ha-manager - Proxmox VE HA Manager |
22653ac8 | 12 | |
49a5e11c | 13 | SYNOPSIS |
22653ac8 DM |
14 | -------- |
15 | ||
16 | include::ha-manager.1-synopsis.adoc[] | |
17 | ||
18 | DESCRIPTION | |
19 | ----------- | |
20 | endif::manvolnum[] | |
21 | ||
22 | ifndef::manvolnum[] | |
23 | High Availability | |
24 | ================= | |
25 | include::attributes.txt[] | |
26 | endif::manvolnum[] | |
5f09af76 DM |
27 | ifdef::wiki[] |
28 | :pve-toplevel: | |
29 | endif::wiki[] | |
b5266e9f DM |
30 | |
31 | Our modern society depends heavily on information provided by | |
32 | computers over the network. Mobile devices amplified that dependency, | |
33 | because people can access the network any time from anywhere. If you | |
34 | provide such services, it is very important that they are available | |
35 | most of the time. | |
36 | ||
37 | We can mathematically define the availability as the ratio of (A) the | |
38 | total time a service is capable of being used during a given interval | |
39 | to (B) the length of the interval. It is normally expressed as a | |
40 | percentage of uptime in a given year. | |
41 | ||
42 | .Availability - Downtime per Year | |
43 | [width="60%",cols="<d,d",options="header"] | |
44 | |=========================================================== | |
45 | |Availability % |Downtime per year | |
46 | |99 |3.65 days | |
47 | |99.9 |8.76 hours | |
48 | |99.99 |52.56 minutes | |
49 | |99.999 |5.26 minutes | |
50 | |99.9999 |31.5 seconds | |
51 | |99.99999 |3.15 seconds | |
52 | |=========================================================== | |
53 | ||
04bde502 DM |
54 | There are several ways to increase availability. The most elegant |
55 | solution is to rewrite your software, so that you can run it on | |
56 | several host at the same time. The software itself need to have a way | |
2af6af05 | 57 | to detect errors and do failover. This is relatively easy if you just |
04bde502 DM |
58 | want to serve read-only web pages. But in general this is complex, and |
59 | sometimes impossible because you cannot modify the software | |
60 | yourself. The following solutions works without modifying the | |
61 | software: | |
62 | ||
8c1189b6 | 63 | * Use reliable ``server'' components |
04bde502 DM |
64 | |
65 | NOTE: Computer components with same functionality can have varying | |
2af6af05 | 66 | reliability numbers, depending on the component quality. Most vendors |
8c1189b6 | 67 | sell components with higher reliability as ``server'' components - |
04bde502 | 68 | usually at higher price. |
b5266e9f DM |
69 | |
70 | * Eliminate single point of failure (redundant components) | |
8c1189b6 FG |
71 | ** use an uninterruptible power supply (UPS) |
72 | ** use redundant power supplies on the main boards | |
73 | ** use ECC-RAM | |
74 | ** use redundant network hardware | |
75 | ** use RAID for local storage | |
76 | ** use distributed, redundant storage for VM data | |
b5266e9f DM |
77 | |
78 | * Reduce downtime | |
8c1189b6 FG |
79 | ** rapidly accessible administrators (24/7) |
80 | ** availability of spare parts (other nodes in a {pve} cluster) | |
81 | ** automatic error detection (provided by `ha-manager`) | |
82 | ** automatic failover (provided by `ha-manager`) | |
b5266e9f | 83 | |
5771d9b0 | 84 | Virtualization environments like {pve} make it much easier to reach |
8c1189b6 | 85 | high availability because they remove the ``hardware'' dependency. They |
04bde502 DM |
86 | also support to setup and use redundant storage and network |
87 | devices. So if one host fail, you can simply start those services on | |
43da8322 DM |
88 | another host within your cluster. |
89 | ||
8c1189b6 | 90 | Even better, {pve} provides a software stack called `ha-manager`, |
43da8322 DM |
91 | which can do that automatically for you. It is able to automatically |
92 | detect errors and do automatic failover. | |
93 | ||
8c1189b6 | 94 | {pve} `ha-manager` works like an ``automated'' administrator. First, you |
43da8322 | 95 | configure what resources (VMs, containers, ...) it should |
8c1189b6 FG |
96 | manage. `ha-manager` then observes correct functionality, and handles |
97 | service failover to another node in case of errors. `ha-manager` can | |
43da8322 DM |
98 | also handle normal user requests which may start, stop, relocate and |
99 | migrate a service. | |
04bde502 DM |
100 | |
101 | But high availability comes at a price. High quality components are | |
102 | more expensive, and making them redundant duplicates the costs at | |
103 | least. Additional spare parts increase costs further. So you should | |
104 | carefully calculate the benefits, and compare with those additional | |
105 | costs. | |
106 | ||
107 | TIP: Increasing availability from 99% to 99.9% is relatively | |
108 | simply. But increasing availability from 99.9999% to 99.99999% is very | |
8c1189b6 | 109 | hard and costly. `ha-manager` has typical error detection and failover |
43da8322 DM |
110 | times of about 2 minutes, so you can get no more than 99.999% |
111 | availability. | |
b5266e9f | 112 | |
5bd515d4 DM |
113 | Requirements |
114 | ------------ | |
3810ae1e | 115 | |
5bd515d4 | 116 | * at least three cluster nodes (to get reliable quorum) |
43da8322 | 117 | |
5bd515d4 | 118 | * shared storage for VMs and containers |
43da8322 | 119 | |
5bd515d4 | 120 | * hardware redundancy (everywhere) |
3810ae1e | 121 | |
5bd515d4 | 122 | * hardware watchdog - if not available we fall back to the |
8c1189b6 | 123 | linux kernel software watchdog (`softdog`) |
3810ae1e | 124 | |
5bd515d4 | 125 | * optional hardware fencing devices |
3810ae1e | 126 | |
3810ae1e | 127 | |
5bd515d4 DM |
128 | Resources |
129 | --------- | |
130 | ||
8c1189b6 FG |
131 | We call the primary management unit handled by `ha-manager` a |
132 | resource. A resource (also called ``service'') is uniquely | |
5bd515d4 | 133 | identified by a service ID (SID), which consists of the resource type |
8c1189b6 FG |
134 | and an type specific ID, e.g.: `vm:100`. That example would be a |
135 | resource of type `vm` (virtual machine) with the ID 100. | |
5bd515d4 DM |
136 | |
137 | For now we have two important resources types - virtual machines and | |
138 | containers. One basic idea here is that we can bundle related software | |
139 | into such VM or container, so there is no need to compose one big | |
8c1189b6 | 140 | service from other services, like it was done with `rgmanager`. In |
5bd515d4 | 141 | general, a HA enabled resource should not depend on other resources. |
3810ae1e | 142 | |
22653ac8 | 143 | |
2b52e195 | 144 | How It Works |
22653ac8 DM |
145 | ------------ |
146 | ||
3810ae1e TL |
147 | This section provides an in detail description of the {PVE} HA-manager |
148 | internals. It describes how the CRM and the LRM work together. | |
149 | ||
150 | To provide High Availability two daemons run on each node: | |
151 | ||
8c1189b6 | 152 | `pve-ha-lrm`:: |
3810ae1e TL |
153 | |
154 | The local resource manager (LRM), it controls the services running on | |
155 | the local node. | |
156 | It reads the requested states for its services from the current manager | |
157 | status file and executes the respective commands. | |
158 | ||
8c1189b6 | 159 | `pve-ha-crm`:: |
3810ae1e TL |
160 | |
161 | The cluster resource manager (CRM), it controls the cluster wide | |
2af6af05 | 162 | actions of the services, processes the LRM results and includes the state |
3810ae1e TL |
163 | machine which controls the state of each service. |
164 | ||
165 | .Locks in the LRM & CRM | |
166 | [NOTE] | |
167 | Locks are provided by our distributed configuration file system (pmxcfs). | |
5771d9b0 TL |
168 | They are used to guarantee that each LRM is active once and working. As a |
169 | LRM only executes actions when it holds its lock we can mark a failed node | |
170 | as fenced if we can acquire its lock. This lets us then recover any failed | |
5eba0743 | 171 | HA services securely without any interference from the now unknown failed node. |
3810ae1e TL |
172 | This all gets supervised by the CRM which holds currently the manager master |
173 | lock. | |
174 | ||
175 | Local Resource Manager | |
176 | ~~~~~~~~~~~~~~~~~~~~~~ | |
177 | ||
8c1189b6 | 178 | The local resource manager (`pve-ha-lrm`) is started as a daemon on |
3810ae1e TL |
179 | boot and waits until the HA cluster is quorate and thus cluster wide |
180 | locks are working. | |
181 | ||
182 | It can be in three states: | |
183 | ||
b8663359 | 184 | wait for agent lock:: |
e1ea726a FG |
185 | |
186 | The LRM waits for our exclusive lock. This is also used as idle state if no | |
187 | service is configured. | |
188 | ||
b8663359 | 189 | active:: |
e1ea726a FG |
190 | |
191 | The LRM holds its exclusive lock and has services configured. | |
192 | ||
b8663359 | 193 | lost agent lock:: |
e1ea726a FG |
194 | |
195 | The LRM lost its lock, this means a failure happened and quorum was lost. | |
3810ae1e TL |
196 | |
197 | After the LRM gets in the active state it reads the manager status | |
8c1189b6 | 198 | file in `/etc/pve/ha/manager_status` and determines the commands it |
2af6af05 | 199 | has to execute for the services it owns. |
3810ae1e | 200 | For each command a worker gets started, this workers are running in |
5eba0743 | 201 | parallel and are limited to at most 4 by default. This default setting |
8c1189b6 | 202 | may be changed through the datacenter configuration key `max_worker`. |
2af6af05 TL |
203 | When finished the worker process gets collected and its result saved for |
204 | the CRM. | |
3810ae1e | 205 | |
5eba0743 | 206 | .Maximum Concurrent Worker Adjustment Tips |
3810ae1e | 207 | [NOTE] |
5eba0743 | 208 | The default value of at most 4 concurrent workers may be unsuited for |
3810ae1e TL |
209 | a specific setup. For example may 4 live migrations happen at the same |
210 | time, which can lead to network congestions with slower networks and/or | |
211 | big (memory wise) services. Ensure that also in the worst case no congestion | |
8c1189b6 | 212 | happens and lower the `max_worker` value if needed. In the contrary, if you |
3810ae1e TL |
213 | have a particularly powerful high end setup you may also want to increase it. |
214 | ||
215 | Each command requested by the CRM is uniquely identifiable by an UID, when | |
216 | the worker finished its result will be processed and written in the LRM | |
8c1189b6 | 217 | status file `/etc/pve/nodes/<nodename>/lrm_status`. There the CRM may collect |
3810ae1e TL |
218 | it and let its state machine - respective the commands output - act on it. |
219 | ||
220 | The actions on each service between CRM and LRM are normally always synced. | |
221 | This means that the CRM requests a state uniquely marked by an UID, the LRM | |
222 | then executes this action *one time* and writes back the result, also | |
223 | identifiable by the same UID. This is needed so that the LRM does not | |
224 | executes an outdated command. | |
8c1189b6 | 225 | With the exception of the `stop` and the `error` command, |
c9aa5d47 | 226 | those two do not depend on the result produced and are executed |
3810ae1e TL |
227 | always in the case of the stopped state and once in the case of |
228 | the error state. | |
229 | ||
230 | .Read the Logs | |
231 | [NOTE] | |
232 | The HA Stack logs every action it makes. This helps to understand what | |
233 | and also why something happens in the cluster. Here its important to see | |
234 | what both daemons, the LRM and the CRM, did. You may use | |
235 | `journalctl -u pve-ha-lrm` on the node(s) where the service is and | |
236 | the same command for the pve-ha-crm on the node which is the current master. | |
237 | ||
238 | Cluster Resource Manager | |
239 | ~~~~~~~~~~~~~~~~~~~~~~~~ | |
22653ac8 | 240 | |
8c1189b6 | 241 | The cluster resource manager (`pve-ha-crm`) starts on each node and |
22653ac8 DM |
242 | waits there for the manager lock, which can only be held by one node |
243 | at a time. The node which successfully acquires the manager lock gets | |
3810ae1e TL |
244 | promoted to the CRM master. |
245 | ||
2af6af05 | 246 | It can be in three states: |
3810ae1e | 247 | |
b8663359 | 248 | wait for agent lock:: |
e1ea726a | 249 | |
97ae300a | 250 | The CRM waits for our exclusive lock. This is also used as idle state if no |
e1ea726a FG |
251 | service is configured |
252 | ||
b8663359 | 253 | active:: |
e1ea726a | 254 | |
97ae300a | 255 | The CRM holds its exclusive lock and has services configured |
e1ea726a | 256 | |
b8663359 | 257 | lost agent lock:: |
e1ea726a | 258 | |
97ae300a | 259 | The CRM lost its lock, this means a failure happened and quorum was lost. |
3810ae1e TL |
260 | |
261 | It main task is to manage the services which are configured to be highly | |
2af6af05 | 262 | available and try to always enforce them to the wanted state, e.g.: a |
3810ae1e | 263 | enabled service will be started if its not running, if it crashes it will |
2af6af05 | 264 | be started again. Thus it dictates the LRM the actions it needs to execute. |
22653ac8 DM |
265 | |
266 | When an node leaves the cluster quorum, its state changes to unknown. | |
267 | If the current CRM then can secure the failed nodes lock, the services | |
268 | will be 'stolen' and restarted on another node. | |
269 | ||
270 | When a cluster member determines that it is no longer in the cluster | |
271 | quorum, the LRM waits for a new quorum to form. As long as there is no | |
272 | quorum the node cannot reset the watchdog. This will trigger a reboot | |
2af6af05 | 273 | after the watchdog then times out, this happens after 60 seconds. |
22653ac8 | 274 | |
2b52e195 | 275 | Configuration |
22653ac8 DM |
276 | ------------- |
277 | ||
2af6af05 | 278 | The HA stack is well integrated in the Proxmox VE API2. So, for |
8c1189b6 | 279 | example, HA can be configured via `ha-manager` or the PVE web |
22653ac8 DM |
280 | interface, which both provide an easy to use tool. |
281 | ||
282 | The resource configuration file can be located at | |
8c1189b6 FG |
283 | `/etc/pve/ha/resources.cfg` and the group configuration file at |
284 | `/etc/pve/ha/groups.cfg`. Use the provided tools to make changes, | |
22653ac8 DM |
285 | there shouldn't be any need to edit them manually. |
286 | ||
3810ae1e TL |
287 | Node Power Status |
288 | ----------------- | |
289 | ||
290 | If a node needs maintenance you should migrate and or relocate all | |
291 | services which are required to run always on another node first. | |
292 | After that you can stop the LRM and CRM services. But note that the | |
293 | watchdog triggers if you stop it with active services. | |
294 | ||
5771d9b0 TL |
295 | Package Updates |
296 | --------------- | |
297 | ||
2af6af05 | 298 | When updating the ha-manager you should do one node after the other, never |
5771d9b0 TL |
299 | all at once for various reasons. First, while we test our software |
300 | thoughtfully, a bug affecting your specific setup cannot totally be ruled out. | |
301 | Upgrading one node after the other and checking the functionality of each node | |
302 | after finishing the update helps to recover from an eventual problems, while | |
303 | updating all could render you in a broken cluster state and is generally not | |
304 | good practice. | |
305 | ||
306 | Also, the {pve} HA stack uses a request acknowledge protocol to perform | |
307 | actions between the cluster and the local resource manager. For restarting, | |
308 | the LRM makes a request to the CRM to freeze all its services. This prevents | |
309 | that they get touched by the Cluster during the short time the LRM is restarting. | |
310 | After that the LRM may safely close the watchdog during a restart. | |
311 | Such a restart happens on a update and as already stated a active master | |
312 | CRM is needed to acknowledge the requests from the LRM, if this is not the case | |
313 | the update process can be too long which, in the worst case, may result in | |
314 | a watchdog reset. | |
315 | ||
2af6af05 | 316 | |
3810ae1e TL |
317 | Fencing |
318 | ------- | |
319 | ||
5eba0743 | 320 | What is Fencing |
3810ae1e TL |
321 | ~~~~~~~~~~~~~~~ |
322 | ||
323 | Fencing secures that on a node failure the dangerous node gets will be rendered | |
324 | unable to do any damage and that no resource runs twice when it gets recovered | |
5771d9b0 TL |
325 | from the failed node. This is a really important task and one of the base |
326 | principles to make a system Highly Available. | |
327 | ||
328 | If a node would not get fenced it would be in an unknown state where it may | |
329 | have still access to shared resources, this is really dangerous! | |
330 | Imagine that every network but the storage one broke, now while not | |
331 | reachable from the public network the VM still runs and writes on the shared | |
332 | storage. If we would not fence the node and just start up this VM on another | |
333 | Node we would get dangerous race conditions, atomicity violations the whole VM | |
334 | could be rendered unusable. The recovery could also simply fail if the storage | |
335 | protects from multiple mounts and thus defeat the purpose of HA. | |
336 | ||
337 | How {pve} Fences | |
338 | ~~~~~~~~~~~~~~~~~ | |
339 | ||
340 | There are different methods to fence a node, for example fence devices which | |
341 | cut off the power from the node or disable their communication completely. | |
342 | ||
343 | Those are often quite expensive and bring additional critical components in | |
344 | a system, because if they fail you cannot recover any service. | |
345 | ||
346 | We thus wanted to integrate a simpler method in the HA Manager first, namely | |
347 | self fencing with watchdogs. | |
348 | ||
349 | Watchdogs are widely used in critical and dependable systems since the | |
350 | beginning of micro controllers, they are often independent and simple | |
351 | integrated circuit which programs can use to watch them. After opening they need to | |
352 | report periodically. If, for whatever reason, a program becomes unable to do | |
353 | so the watchdogs triggers a reset of the whole server. | |
354 | ||
355 | Server motherboards often already include such hardware watchdogs, these need | |
356 | to be configured. If no watchdog is available or configured we fall back to the | |
357 | Linux Kernel softdog while still reliable it is not independent of the servers | |
358 | Hardware and thus has a lower reliability then a hardware watchdog. | |
3810ae1e TL |
359 | |
360 | Configure Hardware Watchdog | |
361 | ~~~~~~~~~~~~~~~~~~~~~~~~~~~ | |
362 | By default all watchdog modules are blocked for security reasons as they are | |
363 | like a loaded gun if not correctly initialized. | |
c9aa5d47 | 364 | If you have a hardware watchdog available remove its kernel module from the |
8c1189b6 | 365 | blacklist, load it with insmod and restart the `watchdog-mux` service or reboot |
c9aa5d47 | 366 | the node. |
3810ae1e | 367 | |
2957ef80 TL |
368 | Recover Fenced Services |
369 | ~~~~~~~~~~~~~~~~~~~~~~~ | |
370 | ||
371 | After a node failed and its fencing was successful we start to recover services | |
372 | to other available nodes and restart them there so that they can provide service | |
373 | again. | |
374 | ||
375 | The selection of the node on which the services gets recovered is influenced | |
376 | by the users group settings, the currently active nodes and their respective | |
377 | active service count. | |
378 | First we build a set out of the intersection between user selected nodes and | |
379 | available nodes. Then the subset with the highest priority of those nodes | |
380 | gets chosen as possible nodes for recovery. We select the node with the | |
381 | currently lowest active service count as a new node for the service. | |
382 | That minimizes the possibility of an overload, which else could cause an | |
383 | unresponsive node and as a result a chain reaction of node failures in the | |
384 | cluster. | |
385 | ||
2b52e195 | 386 | Groups |
22653ac8 DM |
387 | ------ |
388 | ||
389 | A group is a collection of cluster nodes which a service may be bound to. | |
390 | ||
2b52e195 | 391 | Group Settings |
22653ac8 DM |
392 | ~~~~~~~~~~~~~~ |
393 | ||
394 | nodes:: | |
395 | ||
c9aa5d47 TL |
396 | List of group node members where a priority can be given to each node. |
397 | A service bound to this group will run on the nodes with the highest priority | |
398 | available. If more nodes are in the highest priority class the services will | |
399 | get distributed to those node if not already there. The priorities have a | |
400 | relative meaning only. | |
93d2a4f9 | 401 | Example;; |
b352bff4 DM |
402 | You want to run all services from a group on `node1` if possible. If this node |
403 | is not available, you want them to run equally splitted on `node2` and `node3`, and | |
404 | if those fail it should use `node4`. | |
93d2a4f9 TL |
405 | To achieve this you could set the node list to: |
406 | [source,bash] | |
407 | ha-manager groupset mygroup -nodes "node1:2,node2:1,node3:1,node4" | |
22653ac8 DM |
408 | |
409 | restricted:: | |
410 | ||
5eba0743 | 411 | Resources bound to this group may only run on nodes defined by the |
22653ac8 DM |
412 | group. If no group node member is available the resource will be |
413 | placed in the stopped state. | |
93d2a4f9 | 414 | Example;; |
01911cf3 DM |
415 | Lets say a service uses resources only available on `node1` and `node2`, |
416 | so we need to make sure that HA manager does not use other nodes. | |
417 | We need to create a 'restricted' group with said nodes: | |
418 | [source,bash] | |
419 | ha-manager groupset mygroup -nodes "node1,node2" -restricted | |
22653ac8 DM |
420 | |
421 | nofailback:: | |
422 | ||
5eba0743 | 423 | The resource won't automatically fail back when a more preferred node |
22653ac8 | 424 | (re)joins the cluster. |
93d2a4f9 TL |
425 | Examples;; |
426 | * You need to migrate a service to a node which hasn't the highest priority | |
427 | in the group at the moment, to tell the HA manager to not move this service | |
20fa8c22 | 428 | instantly back set the 'nofailback' option and the service will stay on |
345f5fe0 | 429 | the current node. |
93d2a4f9 | 430 | |
345f5fe0 DM |
431 | * A service was fenced and it got recovered to another node. The admin |
432 | repaired the node and brought it up online again but does not want that the | |
93d2a4f9 TL |
433 | recovered services move straight back to the repaired node as he wants to |
434 | first investigate the failure cause and check if it runs stable. He can use | |
345f5fe0 | 435 | the 'nofailback' option to achieve this. |
22653ac8 DM |
436 | |
437 | ||
a3189ad1 TL |
438 | Start Failure Policy |
439 | --------------------- | |
440 | ||
441 | The start failure policy comes in effect if a service failed to start on a | |
442 | node once ore more times. It can be used to configure how often a restart | |
443 | should be triggered on the same node and how often a service should be | |
444 | relocated so that it gets a try to be started on another node. | |
445 | The aim of this policy is to circumvent temporary unavailability of shared | |
446 | resources on a specific node. For example, if a shared storage isn't available | |
447 | on a quorate node anymore, e.g. network problems, but still on other nodes, | |
448 | the relocate policy allows then that the service gets started nonetheless. | |
449 | ||
450 | There are two service start recover policy settings which can be configured | |
22653ac8 DM |
451 | specific for each resource. |
452 | ||
453 | max_restart:: | |
454 | ||
5eba0743 | 455 | Maximum number of tries to restart an failed service on the actual |
22653ac8 DM |
456 | node. The default is set to one. |
457 | ||
458 | max_relocate:: | |
459 | ||
5eba0743 | 460 | Maximum number of tries to relocate the service to a different node. |
22653ac8 DM |
461 | A relocate only happens after the max_restart value is exceeded on the |
462 | actual node. The default is set to one. | |
463 | ||
0abc65b0 | 464 | NOTE: The relocate count state will only reset to zero when the |
22653ac8 DM |
465 | service had at least one successful start. That means if a service is |
466 | re-enabled without fixing the error only the restart policy gets | |
467 | repeated. | |
468 | ||
2b52e195 | 469 | Error Recovery |
22653ac8 DM |
470 | -------------- |
471 | ||
472 | If after all tries the service state could not be recovered it gets | |
473 | placed in an error state. In this state the service won't get touched | |
474 | by the HA stack anymore. To recover from this state you should follow | |
475 | these steps: | |
476 | ||
5eba0743 | 477 | * bring the resource back into a safe and consistent state (e.g., |
22653ac8 DM |
478 | killing its process) |
479 | ||
480 | * disable the ha resource to place it in an stopped state | |
481 | ||
482 | * fix the error which led to this failures | |
483 | ||
484 | * *after* you fixed all errors you may enable the service again | |
485 | ||
486 | ||
2b52e195 | 487 | Service Operations |
22653ac8 DM |
488 | ------------------ |
489 | ||
490 | This are how the basic user-initiated service operations (via | |
8c1189b6 | 491 | `ha-manager`) work. |
22653ac8 DM |
492 | |
493 | enable:: | |
494 | ||
5eba0743 | 495 | The service will be started by the LRM if not already running. |
22653ac8 DM |
496 | |
497 | disable:: | |
498 | ||
5eba0743 | 499 | The service will be stopped by the LRM if running. |
22653ac8 DM |
500 | |
501 | migrate/relocate:: | |
502 | ||
5eba0743 | 503 | The service will be relocated (live) to another node. |
22653ac8 DM |
504 | |
505 | remove:: | |
506 | ||
5eba0743 | 507 | The service will be removed from the HA managed resource list. Its |
22653ac8 DM |
508 | current state will not be touched. |
509 | ||
510 | start/stop:: | |
511 | ||
8c1189b6 FG |
512 | `start` and `stop` commands can be issued to the resource specific tools |
513 | (like `qm` or `pct`), they will forward the request to the | |
514 | `ha-manager` which then will execute the action and set the resulting | |
22653ac8 DM |
515 | service state (enabled, disabled). |
516 | ||
517 | ||
2b52e195 | 518 | Service States |
22653ac8 DM |
519 | -------------- |
520 | ||
521 | stopped:: | |
522 | ||
c9aa5d47 TL |
523 | Service is stopped (confirmed by LRM), if detected running it will get stopped |
524 | again. | |
22653ac8 DM |
525 | |
526 | request_stop:: | |
527 | ||
528 | Service should be stopped. Waiting for confirmation from LRM. | |
529 | ||
530 | started:: | |
531 | ||
532 | Service is active an LRM should start it ASAP if not already running. | |
c9aa5d47 | 533 | If the Service fails and is detected to be not running the LRM restarts it. |
22653ac8 DM |
534 | |
535 | fence:: | |
536 | ||
537 | Wait for node fencing (service node is not inside quorate cluster | |
538 | partition). | |
c9aa5d47 TL |
539 | As soon as node gets fenced successfully the service will be recovered to |
540 | another node, if possible. | |
22653ac8 DM |
541 | |
542 | freeze:: | |
543 | ||
544 | Do not touch the service state. We use this state while we reboot a | |
545 | node, or when we restart the LRM daemon. | |
546 | ||
547 | migrate:: | |
548 | ||
549 | Migrate service (live) to other node. | |
550 | ||
551 | error:: | |
552 | ||
553 | Service disabled because of LRM errors. Needs manual intervention. | |
554 | ||
555 | ||
556 | ifdef::manvolnum[] | |
557 | include::pve-copyright.adoc[] | |
558 | endif::manvolnum[] | |
559 |