1DRBDSETUP(8) System Administration DRBDSETUP(8)
2
6 drbdsetup - Setup tool for DRBD
7
9 drbdsetup {device} disk {lower_dev} {meta_data_dev} {meta_data_index}
10 [-d {size}] [-e {err_handler}] [-f {fencing_policy}] [-b]
11 [-t {disk_timeout}]
12 drbdsetup {device} net [af:] {local_addr} [:port] [af:] {remote_addr}
14 [:port] {protocol} [-c {time}] [-i {time}] [-t {val}]
15 [-S {size}] [-r {size}] [-k {count}] [-e {max_epoch_size}]
16 [-b {max_buffers}] [-m] [-a {hash_alg}] [-x {shared_secret}]
17 [-A {asb-0p-policy}] [-B {asb-1p-policy}]
18 [-C {asb-2p-policy}] [-D] [-R {role-resync-conflict-policy}]
19 [-p {ping_timeout}] [-u {val}] [-d {hash_alg}] [-o] [-n]
20 [-g {congestion_policy}] [-f {val}] [-h {val}]
21 drbdsetup {device} syncer [-a {dev_minor}] [-r {rate}] [-e {extents}]
23 [-v {verify-hash-alg}] [-c {cpu-mask}] [-C {csums-hash-alg}]
24 [-R] [-p {plan_time}] [-s {fill_target}] [-d {delay_target}]
25 [-m {max_rate}] [-n {ond-policy}]
26 drbdsetup {device} disconnect
28 drbdsetup {device} detach [-f]
30 drbdsetup {device} down
32 drbdsetup {device} primary [-f] [-o]
34 drbdsetup {device} secondary
36 drbdsetup {device} verify [-s {start-position}] [-S {stop-position}]
38 drbdsetup {device} invalidate
40 drbdsetup {device} invalidate-remote
42 drbdsetup {device} wait-connect [-t {wfc_timeout}]
44 [-d {degr_wfc_timeout}] [-o {outdated_wfc_timeout}] [-w]
45 drbdsetup {device} wait-sync [-t {wfc_timeout}] [-d {degr_wfc_timeout}]
47 [-o {outdated_wfc_timeout}] [-w]
48 drbdsetup {device} role
50 drbdsetup {device} cstate
52 drbdsetup {device} dstate
54 drbdsetup {device} status
56 drbdsetup {device} resize [-d {size}] [-f {assume-peer-has-space}]
58 [-c {assume-clean}]
59 drbdsetup {device} check-resize
61 drbdsetup {device} pause-sync
63 drbdsetup {device} resume-sync
65 drbdsetup {device} outdate
67 drbdsetup {device} show-gi
69 drbdsetup {device} get-gi
71 drbdsetup {device} show
73 drbdsetup {device} suspend-io
75 drbdsetup {device} resume-io
77 drbdsetup {device} events [-u] [-a]
79 drbdsetup {device} new-current-uuid [-c]
81
83 drbdsetup is used to associate DRBD devices with their backing block
84 devices, to set up DRBD device pairs to mirror their backing block
85 devices, and to inspect the configuration of running DRBD devices.
86
88 drbdsetup is a low level tool of the DRBD program suite. It is used by
89 the data disk and drbd scripts to communicate with the device driver.
90
92 Each drbdsetup sub-command might require arguments and bring its own
93 set of options. All values have default units which might be overruled
94 by K, M or G. These units are defined in the usual way (e.g. K = 2^10 =
95 1024).
96 Common options
98 All drbdsetup sub-commands accept these two options
99 --create-device
101 In case the specified DRBD device (minor number) does not exist
102 yet, create it implicitly.
103 --set-defaults
105 When --set-defaults is given on the command line, all options of
106 the invoked sub-command that are not explicitly set are reset to
107 their default values.
108 disk
110 Associates device with lower_device to store its data blocks on. The -d
111 (or --disk-size) should only be used if you wish not to use as much as
112 possible from the backing block devices. If you do not use -d, the
113 device is only ready for use as soon as it was connected to its peer
114 once. (See the net command.)
115 -d, --disk-size size
117 You can override DRBD's size determination method with this option.
118 If you need to use the device before it was ever connected to its
119 peer, use this option to pass the size of the DRBD device to the
120 driver. Default unit is sectors (1s = 512 bytes).
121 If you use the size parameter in drbd.conf, we strongly recommend
123 to add an explicit unit postfix. drbdadm and drbdsetup used to have
124 mismatching default units.
125 -e, --on-io-error err_handler
127 If the driver of the lower_device reports an error to DRBD, DRBD
128 will mark the disk as inconsistent, call a helper program, or
129 detach the device from its backing storage and perform all further
130 IO by requesting it from the peer. The valid err_handlers are:
131 pass_on, call-local-io-error and detach.
132 -f, --fencing fencing_policy
134 Under fencing we understand preventive measures to avoid situations
135 where both nodes are primary and disconnected (AKA split brain).
136 Valid fencing policies are:
138 dont-care
140 This is the default policy. No fencing actions are done.
141 resource-only
143 If a node becomes a disconnected primary, it tries to outdate
144 the peer's disk. This is done by calling the fence-peer
145 handler. The handler is supposed to reach the other node over
146 alternative communication paths and call 'drbdadm outdate res'
147 there.
148 resource-and-stonith
150 If a node becomes a disconnected primary, it freezes all its IO
151 operations and calls its fence-peer handler. The fence-peer
152 handler is supposed to reach the peer over alternative
153 communication paths and call 'drbdadm outdate res' there. In
154 case it cannot reach the peer, it should stonith the peer. IO
155 is resumed as soon as the situation is resolved. In case your
156 handler fails, you can resume IO with the resume-io command.
157 -b, --use-bmbv
159 In case the backing storage's driver has a merge_bvec_fn()
160 function, DRBD has to pretend that it can only process IO requests
161 in units not larger than 4 KiB. (At time of writing the only known
162 drivers which have such a function are: md (software raid driver),
163 dm (device mapper - LVM) and DRBD itself)
164 To get best performance out of DRBD on top of software raid (or any
166 other driver with a merge_bvec_fn() function) you might enable this
167 option, if you know for sure that the merge_bvec_fn() function will
168 deliver the same results on all nodes of your cluster. I.e. the
169 physical disks of the software raid are exactly of the same type.
170 USE THIS OPTION ONLY IF YOU KNOW WHAT YOU ARE DOING.
171 -a, --no-disk-barrier, -i, --no-disk-flushes, -D, --no-disk-drain
173 DRBD has four implementations to express write-after-write
174 dependencies to its backing storage device. DRBD will use the first
175 method that is supported by the backing storage device and that is
176 not disabled by the user.
177 When selecting the method you should not only base your decision on
179 the measurable performance. In case your backing storage device has
180 a volatile write cache (plain disks, RAID of plain disks) you
181 should use one of the first two. In case your backing storage
182 device has battery-backed write cache you may go with option 3.
183 Option 4 (disable everything, use "none") is dangerous on most IO
184 stacks, may result in write-reordering, and if so, can
185 theoretically be the reason for data corruption, or disturb the
186 DRBD protocol, causing spurious disconnect/reconnect cycles. Do
187 not use no-disk-drain.
188 Unfortunately device mapper (LVM) might not support barriers.
190 The letter after "wo:" in /proc/drbd indicates with method is
192 currently in use for a device: b, f, d, n. The implementations:
193 barrier
195 The first requires that the driver of the backing storage
196 device support barriers (called 'tagged command queuing' in
197 SCSI and 'native command queuing' in SATA speak). The use of
198 this method can be disabled by the --no-disk-barrier option.
199 Note: Since Linux-2.6.36 (or RHEL's 2.6.32) this method is
200 disabled.
201 flush
203 The second requires that the backing device support disk
204 flushes (called 'force unit access' in the drive vendors
205 speak). The use of this method can be disabled using the
206 --no-disk-flushes option.
207 drain
209 The third method is simply to let write requests drain before
210 write requests of a new reordering domain are issued. That was
211 the only implementation before 8.0.9.
212 none
214 The fourth method is to not express write-after-write
215 dependencies to the backing store at all, by also specifying
216 --no-disk-drain. This is dangerous on most IO stacks, may
217 result in write-reordering, and if so, can theoretically be the
218 reason for data corruption, or disturb the DRBD protocol,
219 causing spurious disconnect/reconnect cycles. Do not use
220 --no-disk-drain.
221 -m, --no-md-flushes
223 Disables the use of disk flushes and barrier BIOs when accessing
224 the meta data device. See the notes on --no-disk-flushes.
225 -s, --max-bio-bvecs
227 In some special circumstances the device mapper stack manages to
228 pass BIOs to DRBD that violate the constraints that are set forth
229 by DRBD's merge_bvec() function and which have more than one bvec.
230 A known example is: phys-disk -> DRBD -> LVM -> Xen -> missaligned
231 partition (63) -> DomU FS. Then you might see "bio would need to,
232 but cannot, be split:" in the Dom0's kernel log.
233 The best workaround is to proper align the partition within the VM
235 (E.g. start it at sector 1024). That costs 480 KiB of storage.
236 Unfortunately the default of most Linux partitioning tools is to
237 start the first partition at an odd number (63). Therefore most
238 distributions install helpers for virtual linux machines will end
239 up with missaligned partitions. The second best workaround is to
240 limit DRBD's max bvecs per BIO (i.e., the max-bio-bvecs option) to
241 1, but that might cost performance.
242 The default value of max-bio-bvecs is 0, which means that there is
244 no user imposed limitation.
245 -t, --disk-timeout disk_timeout
247 If the driver of the lower_device does not finish an IO request
248 within disk_timeout, DRBD considers the disk as failed. If DRBD is
249 connected to a remote host, it will reissue local pending IO
250 requests to the peer, and ship all new IO requests to the peer
251 only. The disk state advances to diskless, as soon as the backing
252 block device has finished all IO requests.
253 The default value of disk-timeout is 0, which means that no timeout
255 is enforced. The default unit is 100ms. This option is available
256 since 8.3.12.
257 net
259 Sets up the device to listen on af:local_addr:port for incoming
260 connections and to try to connect to af:remote_addr:port. If port is
261 omitted, 7788 is used as default. If af is omitted ipv4 gets used.
262 Other supported address families are ipv6, ssocks for Dolphin
263 Interconnect Solutions' "super sockets" and sdp for Sockets Direct
264 Protocol (Infiniband).
265 On the TCP/IP link the specified protocol is used. Valid protocol
267 specifiers are A, B, and C.
268 Protocol A: write IO is reported as completed, if it has reached local
270 disk and local TCP send buffer.
271 Protocol B: write IO is reported as completed, if it has reached local
273 disk and remote buffer cache.
274 Protocol C: write IO is reported as completed, if it has reached both
276 local and remote disk.
277 -c, --connect-int time
279 In case it is not possible to connect to the remote DRBD device
280 immediately, DRBD keeps on trying to connect. With this option you
281 can set the time between two retries. The default value is 10
282 seconds, the unit is 1 second.
283 -i, --ping-int time
285 If the TCP/IP connection linking a DRBD device pair is idle for
286 more than time seconds, DRBD will generate a keep-alive packet to
287 check if its partner is still alive. The default value is 10
288 seconds, the unit is 1 second.
289 -t, --timeout val
291 If the partner node fails to send an expected response packet
292 within val tenths of a second, the partner node is considered dead
293 and therefore the TCP/IP connection is abandoned. The default value
294 is 60 (= 6 seconds).
295 -S, --sndbuf-size size
297 The socket send buffer is used to store packets sent to the
298 secondary node, which are not yet acknowledged (from a network
299 point of view) by the secondary node. When using protocol A, it
300 might be necessary to increase the size of this data structure in
301 order to increase asynchronicity between primary and secondary
302 nodes. But keep in mind that more asynchronicity is synonymous with
303 more data loss in the case of a primary node failure. Since 8.0.13
304 resp. 8.2.7 setting the size value to 0 means that the kernel
305 should autotune this. The default size is 0, i.e. autotune.
306 -r, --rcvbuf-size size
308 Packets received from the network are stored in the socket receive
309 buffer first. From there they are consumed by DRBD. Before 8.3.2
310 the receive buffer's size was always set to the size of the socket
311 send buffer. Since 8.3.2 they can be tuned independently. A value
312 of 0 means that the kernel should autotune this. The default size
313 is 0, i.e. autotune.
314 -k, --ko-count count
316 In case the secondary node fails to complete a single write request
317 for count times the timeout, it is expelled from the cluster, i.e.
318 the primary node goes into StandAlone mode. To disable this
319 feature, you should explicitly set it to 0; defaults may change
320 between versions.
321 -e, --max-epoch-size val
323 With this option the maximal number of write requests between two
324 barriers is limited. Typically set to the same as --max-buffers, or
325 the allowed maximum. Values smaller than 10 can lead to degraded
326 performance. The default value is 2048.
327 -b, --max-buffers val
329 With this option the maximal number of buffer pages allocated by
330 DRBD's receiver thread is limited. Typically set to the same as
331 --max-epoch-size. Small values could lead to degraded performance.
332 The default value is 2048, the minimum 32. Increase this if you
333 cannot saturate the IO backend of the receiving side during linear
334 write or during resync while otherwise idle.
335 See also drbd.conf(5)
337 -u, --unplug-watermark val
339 This setting has no effect with recent kernels that use explicit
340 on-stack plugging (upstream Linux kernel 2.6.39, distributions may
341 have backported).
342 When the number of pending write requests on the standby
344 (secondary) node exceeds the unplug-watermark, we trigger the
345 request processing of our backing storage device. Some storage
346 controllers deliver better performance with small values, others
347 deliver best performance when the value is set to the same value as
348 max-buffers, yet others don't feel much effect at all. Minimum 16,
349 default 128, maximum 131072.
350 -m, --allow-two-primaries
352 With this option set you may assign primary role to both nodes. You
353 only should use this option if you use a shared storage file system
354 on top of DRBD. At the time of writing the only ones are: OCFS2 and
355 GFS. If you use this option with any other file system, you are
356 going to crash your nodes and to corrupt your data!
357 -a, --cram-hmac-alg alg
359 You need to specify the HMAC algorithm to enable peer
360 authentication at all. You are strongly encouraged to use peer
361 authentication. The HMAC algorithm will be used for the challenge
362 response authentication of the peer. You may specify any digest
363 algorithm that is named in /proc/crypto.
364 -x, --shared-secret secret
366 The shared secret used in peer authentication. May be up to 64
367 characters.
368 -A, --after-sb-0pri asb-0p-policy
370 possible policies are:
371 disconnect
373 No automatic resynchronization, simply disconnect.
374 discard-younger-primary
376 Auto sync from the node that was primary before the split-brain
377 situation occurred.
378 discard-older-primary
380 Auto sync from the node that became primary as second during
381 the split-brain situation.
382 discard-zero-changes
384 In case one node did not write anything since the split brain
385 became evident, sync from the node that wrote something to the
386 node that did not write anything. In case none wrote anything
387 this policy uses a random decision to perform a "resync" of 0
388 blocks. In case both have written something this policy
389 disconnects the nodes.
390 discard-least-changes
392 Auto sync from the node that touched more blocks during the
393 split brain situation.
394 discard-node-NODENAME
396 Auto sync to the named node.
397 -B, --after-sb-1pri asb-1p-policy
399 possible policies are:
400 disconnect
402 No automatic resynchronization, simply disconnect.
403 consensus
405 Discard the version of the secondary if the outcome of the
406 after-sb-0pri algorithm would also destroy the current
407 secondary's data. Otherwise disconnect.
408 discard-secondary
410 Discard the secondary's version.
411 call-pri-lost-after-sb
413 Always honor the outcome of the after-sb-0pri algorithm. In
414 case it decides the current secondary has the correct data,
415 call the pri-lost-after-sb on the current primary.
416 violently-as0p
418 Always honor the outcome of the after-sb-0pri algorithm. In
419 case it decides the current secondary has the correct data,
420 accept a possible instantaneous change of the primary's data.
421 -C, --after-sb-2pri asb-2p-policy
423 possible policies are:
424 disconnect
426 No automatic resynchronization, simply disconnect.
427 call-pri-lost-after-sb
429 Always honor the outcome of the after-sb-0pri algorithm. In
430 case it decides the current secondary has the right data, call
431 the pri-lost-after-sb on the current primary.
432 violently-as0p
434 Always honor the outcome of the after-sb-0pri algorithm. In
435 case it decides the current secondary has the right data,
436 accept a possible instantaneous change of the primary's data.
437 -P, --always-asbp
439 Normally the automatic after-split-brain policies are only used if
440 current states of the UUIDs do not indicate the presence of a third
441 node.
442 With this option you request that the automatic after-split-brain
444 policies are used as long as the data sets of the nodes are somehow
445 related. This might cause a full sync, if the UUIDs indicate the
446 presence of a third node. (Or double faults have led to strange
447 UUID sets.)
448 -R, --rr-conflict role-resync-conflict-policy
450 This option sets DRBD's behavior when DRBD deduces from its meta
451 data that a resynchronization is needed, and the SyncTarget node is
452 already primary. The possible settings are: disconnect,
453 call-pri-lost and violently. While disconnect speaks for itself,
454 with the call-pri-lost setting the pri-lost handler is called which
455 is expected to either change the role of the node to secondary, or
456 remove the node from the cluster. The default is disconnect.
457 With the violently setting you allow DRBD to force a primary node
459 into SyncTarget state. This means that the data exposed by DRBD
460 changes to the SyncSource's version of the data instantaneously.
461 USE THIS OPTION ONLY IF YOU KNOW WHAT YOU ARE DOING.
462 -d, --data-integrity-alg hash_alg
464 DRBD can ensure the data integrity of the user's data on the
465 network by comparing hash values. Normally this is ensured by the
466 16 bit checksums in the headers of TCP/IP packets. This option can
467 be set to any of the kernel's data digest algorithms. In a typical
468 kernel configuration you should have at least one of md5, sha1, and
469 crc32c available. By default this is not enabled.
470 See also the notes on data integrity on the drbd.conf manpage.
472 -o, --no-tcp-cork
474 DRBD usually uses the TCP socket option TCP_CORK to hint to the
475 network stack when it can expect more data, and when it should
476 flush out what it has in its send queue. There is at least one
477 network stack that performs worse when one uses this hinting
478 method. Therefore we introduced this option, which disable the
479 setting and clearing of the TCP_CORK socket option by DRBD.
480 -p, --ping-timeout ping_timeout
482 The time the peer has to answer to a keep-alive packet. In case the
483 peer's reply is not received within this time period, it is
484 considered dead. The default unit is tenths of a second, the
485 default value is 5 (for half a second).
486 -D, --discard-my-data
488 Use this option to manually recover from a split-brain situation.
489 In case you do not have any automatic after-split-brain policies
490 selected, the nodes refuse to connect. By passing this option you
491 make this node a sync target immediately after successful connect.
492 -n, --dry-run
494 Causes DRBD to abort the connection process after the resync
495 handshake, i.e. no resync gets performed. You can find out which
496 resync DRBD would perform by looking at the kernel's log file.
497 -g, --on-congestion congestion_policy, -f, --congestion-fill
499 fill_threshold, -h, --congestion-extents active_extents_threshold
500 By default DRBD blocks when the available TCP send queue becomes
501 full. That means it will slow down the application that generates
502 the write requests that cause DRBD to send more data down that TCP
503 connection.
504 When DRBD is deployed with DRBD-proxy it might be more desirable
506 that DRBD goes into AHEAD/BEHIND mode shortly before the send queue
507 becomes full. In AHEAD/BEHIND mode DRBD does no longer replicate
508 data, but still keeps the connection open.
509 The advantage of the AHEAD/BEHIND mode is that the application is
511 not slowed down, even if DRBD-proxy's buffer is not sufficient to
512 buffer all write requests. The downside is that the peer node falls
513 behind, and that a resync will be necessary to bring it back into
514 sync. During that resync the peer node will have an inconsistent
515 disk.
516 Available congestion_policys are block and pull-ahead. The default
518 is block. Fill_threshold might be in the range of 0 to 10GiBytes.
519 The default is 0 which disables the check.
520 Active_extents_threshold has the same limits as al-extents.
521 The AHEAD/BEHIND mode and its settings are available since DRBD
523 8.3.10.
524 syncer
526 Changes the synchronization daemon parameters of device at runtime.
527 -r, --rate rate
529 To ensure smooth operation of the application on top of DRBD, it is
530 possible to limit the bandwidth that may be used by background
531 synchronization. The default is 250 KiB/sec, the default unit is
532 KiB/sec.
533 -a, --after minor
535 Start resync on this device only if the device with minor is
536 already in connected state. Otherwise this device waits in
537 SyncPause state.
538 -e, --al-extents extents
540 DRBD automatically performs hot area detection. With this parameter
541 you control how big the hot area (=active set) can get. Each extent
542 marks 4M of the backing storage. In case a primary node leaves the
543 cluster unexpectedly, the areas covered by the active set must be
544 resynced upon rejoining of the failed node. The data structure is
545 stored in the meta-data area, therefore each change of the active
546 set is a write operation to the meta-data device. A higher number
547 of extents gives longer resync times but less updates to the
548 meta-data. The default number of extents is 127. (Minimum: 7,
549 Maximum: 3843)
550 -v, --verify-alg hash-alg
552 During online verification (as initiated by the verify
553 sub-command), rather than doing a bit-wise comparison, DRBD applies
554 a hash function to the contents of every block being verified, and
555 compares that hash with the peer. This option defines the hash
556 algorithm being used for that purpose. It can be set to any of the
557 kernel's data digest algorithms. In a typical kernel configuration
558 you should have at least one of md5, sha1, and crc32c available. By
559 default this is not enabled; you must set this option explicitly in
560 order to be able to use on-line device verification.
561 See also the notes on data integrity on the drbd.conf manpage.
563 -c, --cpu-mask cpu-mask
565 Sets the cpu-affinity-mask for DRBD's kernel threads of this
566 device. The default value of cpu-mask is 0, which means that DRBD's
567 kernel threads should be spread over all CPUs of the machine. This
568 value must be given in hexadecimal notation. If it is too big it
569 will be truncated.
570 -C, --csums-alg hash-alg
572 A resync process sends all marked data blocks form the source to
573 the destination node, as long as no csums-alg is given. When one is
574 specified the resync process exchanges hash values of all marked
575 blocks first, and sends only those data blocks over, that have
576 different hash values.
577 This setting is useful for DRBD setups with low bandwidth links.
579 During the restart of a crashed primary node, all blocks covered by
580 the activity log are marked for resync. But a large part of those
581 will actually be still in sync, therefore using csums-alg will
582 lower the required bandwidth in exchange for CPU cycles.
583 -R, --use-rle
585 During resync-handshake, the dirty-bitmaps of the nodes are
586 exchanged and merged (using bit-or), so the nodes will have the
587 same understanding of which blocks are dirty. On large devices, the
588 fine grained dirty-bitmap can become large as well, and the bitmap
589 exchange can take quite some time on low-bandwidth links.
590 Because the bitmap typically contains compact areas where all bits
592 are unset (clean) or set (dirty), a simple run-length encoding
593 scheme can considerably reduce the network traffic necessary for
594 the bitmap exchange.
595 For backward compatibilty reasons, and because on fast links this
597 possibly does not improve transfer time but consumes cpu cycles,
598 this defaults to off.
599 Introduced in 8.3.2.
601 -p, --c-plan-ahead plan_time, -s, --c-fill-target fill_target, -d,
603 --c-delay-target delay_target, -M, --c-max-rate max_rate
604 The dynamic resync speed controller gets enabled with setting
605 plan_time to a positive value. It aims to fill the buffers along
606 the data path with either a constant amount of data fill_target, or
607 aims to have a constant delay time of delay_target along the path.
608 The controller has an upper bound of max_rate.
609 By plan_time the agility of the controller is configured. Higher
611 values yield for slower/lower responses of the controller to
612 deviation from the target value. It should be at least 5 times RTT.
613 For regular data paths a fill_target in the area of 4k to 100k is
614 appropriate. For a setup that contains drbd-proxy it is advisable
615 to use delay_target instead. Only when fill_target is set to 0 the
616 controller will use delay_target. 5 times RTT is a reasonable
617 starting value. Max_rate should be set to the bandwidth available
618 between the DRBD-hosts and the machines hosting DRBD-proxy, or to
619 the available disk-bandwidth.
620 The default value of plan_time is 0, the default unit is 0.1
622 seconds. Fill_target has 0 and sectors as default unit.
623 Delay_target has 1 (100ms) and 0.1 as default unit. Max_rate has
624 10240 (100MiB/s) and KiB/s as default unit.
625 -m, --c-min-rate min_rate
627 We track the disk IO rate caused by the resync, so we can detect
628 non-resync IO on the lower level device. If the lower level device
629 seems to be busy, and the current resync rate is above min_rate, we
630 throttle the resync.
631 The default value of min_rate is 4M, the default unit is k. If you
633 want to not throttle at all, set it to zero, if you want to
634 throttle always, set it to one.
635 -n, --on-no-data-accessible ond-policy
637 This setting controls what happens to IO requests on a degraded,
638 disk less node (I.e. no data store is reachable). The available
639 policies are io-error and suspend-io.
640 If ond-policy is set to suspend-io you can either resume IO by
642 attaching/connecting the last lost data storage, or by the drbdadm
643 resume-io res command. The latter will result in IO errors of
644 course.
645 The default is io-error. This setting is available since DRBD
647 8.3.9.
648 primary
650 Sets the device into primary role. This means that applications (e.g. a
651 file system) may open the device for read and write access. Data
652 written to the device in primary role are mirrored to the device in
653 secondary role.
654 Normally it is not possible to set both devices of a connected DRBD
656 device pair to primary role. By using the --allow-two-primaries option,
657 you override this behavior and instruct DRBD to allow two primaries.
658 -o, --overwrite-data-of-peer
660 Alias for --force.
661 -f, --force
663 Becoming primary fails if the local replica is not up-to-date. I.e.
664 when it is inconsistent, outdated of consistent. By using this
665 option you can force it into primary role anyway. USE THIS OPTION
666 ONLY IF YOU KNOW WHAT YOU ARE DOING.
667 secondary
669 Brings the device into secondary role. This operation fails as long as
670 at least one application (or file system) has opened the device.
671 It is possible that both devices of a connected DRBD device pair are
673 secondary.
674 verify
676 This initiates on-line device verification. During on-line
677 verification, the contents of every block on the local node are
678 compared to those on the peer node. Device verification progress can be
679 monitored via /proc/drbd. Any blocks whose content differs from that of
680 the corresponding block on the peer node will be marked out-of-sync in
681 DRBD's on-disk bitmap; they are not brought back in sync automatically.
682 To do that, simply disconnect and reconnect the resource.
683 If on-line verification is already in progress (and this node is
685 "VerifyS"), this command silently "succeeds". In this case, any
686 start-sector (see below) will be ignored, and any stop-sector (see
687 below) will be honored. This can be used to stop a running verify, or
688 to update/shorten/extend the coverage of the currently running verify.
689 This command will fail if the device is not part of a connected device
691 pair.
692 See also the notes on data integrity on the drbd.conf manpage.
694 -s, --start start-sector
696 Since version 8.3.2, on-line verification should resume from the
697 last position after connection loss. It may also be started from an
698 arbitrary position by setting this option. If you had reached some
699 stop-sector before, and you do not specify an explicit
700 start-sector, verify should resume from the previous stop-sector.
701 Default unit is sectors. You may also specify a unit explicitly.
703 The start-sector will be rounded down to a multiple of 8 sectors
704 (4kB).
705 -S, --stop stop-sector
707 Since version 8.3.14, on-line verification can be stopped before it
708 reaches end-of-device. This can be
709 Default unit is sectors. You may also specify a unit explicitly.
711 The stop-sector may be updated by issuing an additional drbdsetup
712 verify command on the same node while the verify is running.
713 invalidate
715 This forces the local device of a pair of connected DRBD devices into
716 SyncTarget state, which means that all data blocks of the device are
717 copied over from the peer.
718 This command will fail if the device is not either part of a connected
720 device pair, or disconnected Secondary.
721 invalidate-remote
723 This forces the local device of a pair of connected DRBD devices into
724 SyncSource state, which means that all data blocks of the device are
725 copied to the peer.
726 On a disconnected Primary device, this will set all bits in the out of
728 sync bitmap. As a side affect this suspends updates to the on disk
729 activity log. Updates to the on disk activity log resume automatically
730 when necessary.
731 wait-connect
733 Returns as soon as the device can communicate with its partner device.
734 -t, --wfc-timeout wfc_timeout, -d, --degr-wfc-timeout degr_wfc_timeout,
736 -o, --outdated-wfc-timeout outdated_wfc_timeout, -w, --wait-after-sb
737 This command will fail if the device cannot communicate with its
738 partner for timeout seconds. If the peer was working before this
739 node was rebooted, the wfc_timeout is used. If the peer was already
740 down before this node was rebooted, the degr_wfc_timeout is used.
741 If the peer was successfully outdated before this node was rebooted
742 the outdated_wfc_timeout is used. The default value for all those
743 timeout values is 0 which means to wait forever. In case the
744 connection status goes down to StandAlone because the peer appeared
745 but the devices had a split brain situation, the default for the
746 command is to terminate. You can change this behavior with the
747 --wait-after-sb option.
748 wait-sync
750 Returns as soon as the device leaves any synchronization into connected
751 state. The options are the same as with the wait-connect command.
752 disconnect
754 Removes the information set by the net command from the device. This
755 means that the device goes into unconnected state and will no longer
756 listen for incoming connections.
757 detach
759 Removes the information set by the disk command from the device. This
760 means that the device is detached from its backing storage device.
761 -f, --force
763 A regular detach returns after the disk state finally reached
764 diskless. As a consequence detaching from a frozen backing block
765 device never terminates.
766 On the other hand A forced detach returns immediately. It allows
768 you to detach DRBD from a frozen backing block device. Please note
769 that the disk will be marked as failed until all pending IO
770 requests where finished by the backing block device.
771 down
773 Removes all configuration information from the device and forces it
774 back to unconfigured state.
775 role
777 Shows the current roles of the device and its peer, as local/peer.
778 state
780 Deprecated alias for "role"
781 cstate
783 Shows the current connection state of the device.
784 dstate
786 Shows the current states of the backing storage devices, as local/peer.
787 status
789 Shows the current status of the device in XML-like format. Example
790 output:
791 <resource minor="0" name="s0" cs="SyncTarget" st1="Secondary" st2="Secondary"
793 ds1="Inconsistent" ds2="UpToDate" resynced_precent="5.9" />
794 resize
797 This causes DRBD to reexamine the size of the device's backing storage
798 device. To actually do online growing you need to extend the backing
799 storages on both devices and call the resize command on one of your
800 nodes.
801 The --assume-peer-has-space allows you to resize a device which is
803 currently not connected to the peer. Use with care, since if you do not
804 resize the peer's disk as well, further connect attempts of the two
805 will fail.
806 When the --assume-clean option is given DRBD will skip the resync of
808 the new storage. Only do this if you know that the new storage was
809 initialized to the same content by other means.
810 check-resize
812 To enable DRBD to detect offline resizing of backing devices this
813 command may be used to record the current size of backing devices. The
814 size is stored in files in /var/lib/drbd/ named drbd-minor-??.lkbd
815 This command is called by drbdadm resize res after drbdsetup device
817 resize returned.
818 pause-sync
820 Temporarily suspend an ongoing resynchronization by setting the local
821 pause flag. Resync only progresses if neither the local nor the remote
822 pause flag is set. It might be desirable to postpone DRBD's
823 resynchronization after eventual resynchronization of the backing
824 storage's RAID setup.
825 resume-sync
827 Unset the local sync pause flag.
828 outdate
830 Mark the data on the local backing storage as outdated. An outdated
831 device refuses to become primary. This is used in conjunction with
832 fencing and by the peer's fence-peer handler.
833 show-gi
835 Displays the device's data generation identifiers verbosely.
836 get-gi
838 Displays the device's data generation identifiers.
839 show
841 Shows all available configuration information of the device.
842 suspend-io
844 This command is of no apparent use and just provided for the sake of
845 completeness.
846 resume-io
848 If the fence-peer handler fails to stonith the peer node, and your
849 fencing policy is set to resource-and-stonith, you can unfreeze IO
850 operations with this command.
851 events
853 Displays every state change of DRBD and all calls to helper programs.
854 This might be used to get notified of DRBD's state changes by piping
855 the output to another program.
856 -a, --all-devices
858 Display the events of all DRBD minors.
859 -u, --unfiltered
861 This is a debugging aid that displays the content of all received
862 netlink messages.
863 new-current-uuid
865 Generates a new current UUID and rotates all other UUID values. This
866 has at least two use cases, namely to skip the initial sync, and to
867 reduce network bandwidth when starting in a single node configuration
868 and then later (re-)integrating a remote site.
869 Available option:
871 -c, --clear-bitmap
873 Clears the sync bitmap in addition to generating a new current
874 UUID.
875 This can be used to skip the initial sync, if you want to start from
877 scratch. This use-case does only work on "Just Created" meta data.
878 Necessary steps:
879 1. On both nodes, initialize meta data and configure the device.
881 drbdadm -- --force create-md res
883 2. They need to do the initial handshake, so they know their sizes.
885 drbdadm up res
887 3. They are now Connected Secondary/Secondary
889 Inconsistent/Inconsistent. Generate a new current-uuid and clear
890 the dirty bitmap.
891 drbdadm -- --clear-bitmap new-current-uuid res
893 4. They are now Connected Secondary/Secondary UpToDate/UpToDate. Make
895 one side primary and create a file system.
896 drbdadm primary res
898 mkfs -t fs-type $(drbdadm sh-dev res)
900 One obvious side-effect is that the replica is full of old garbage
902 (unless you made them identical using other means), so any
903 online-verify is expected to find any number of out-of-sync blocks.
904 You must not use this on pre-existing data! Even though it may appear
906 to work at first glance, once you switch to the other node, your data
907 is toast, as it never got replicated. So do not leave out the mkfs (or
908 equivalent).
909 This can also be used to shorten the initial resync of a cluster where
911 the second node is added after the first node is gone into production,
912 by means of disk shipping. This use-case works on disconnected devices
913 only, the device may be in primary or secondary role.
914 The necessary steps on the current active server are:
916 1. drbdsetup device new-current-uuid --clear-bitmap
918 2. Take the copy of the current active server. E.g. by pulling a disk
920 out of the RAID1 controller, or by copying with dd. You need to
921 copy the actual data, and the meta data.
922 3. drbdsetup device new-current-uuid
924 Now add the disk to the new secondary node, and join it to the cluster.
926 You will get a resync of that parts that were changed since the first
927 call to drbdsetup in step 1.
928
930 For examples, please have a look at the DRBD User's Guide[1].
931
933 This document was revised for version 8.3.2 of the DRBD distribution.
934
936 Written by Philipp Reisner <philipp.reisner@linbit.com> and Lars
937 Ellenberg <lars.ellenberg@linbit.com>
938
940 Report bugs to <drbd-user@lists.linbit.com>.
941
943 Copyright 2001-2008 LINBIT Information Technologies, Philipp Reisner,
944 Lars Ellenberg. This is free software; see the source for copying
945 conditions. There is NO warranty; not even for MERCHANTABILITY or
946 FITNESS FOR A PARTICULAR PURPOSE.
947
949 drbd.conf(5), drbd(8), drbddisk(8), drbdadm(8), DRBD User's Guide[1],
950 DRBD web site[2]
951
953 1. DRBD User's Guide
954 http://www.drbd.org/users-guide/
955 2. DRBD web site
957 http://www.drbd.org/
958DRBD 8.3.2 5 Dec 2008 DRBDSETUP(8)