1OVSDB(7) Open vSwitch OVSDB(7)
2
6 ovsdb - Open vSwitch Database (Overview)
7
9 OVSDB, the Open vSwitch Database, is a network-accessible database sys‐
10 tem. Schemas in OVSDB specify the tables in a database and their col‐
11 umns’ types and can include data, uniqueness, and referential integrity
12 constraints. OVSDB offers atomic, consistent, isolated, durable trans‐
13 actions. RFC 7047 specifies the JSON-RPC based protocol that OVSDB
14 clients and servers use to communicate.
15 The OVSDB protocol is well suited for state synchronization because it
17 allows each client to monitor the contents of a whole database or a
18 subset of it. Whenever a monitored portion of the database changes,
19 the server tells the client what rows were added or modified (including
20 the new contents) or deleted. Thus, OVSDB clients can easily keep
21 track of the newest contents of any part of the database.
22 While OVSDB is general-purpose and not particularly specialized for use
24 with Open vSwitch, Open vSwitch does use it for multiple purposes. The
25 leading use of OVSDB is for configuring and monitoring ovs-vswitchd(8),
26 the Open vSwitch switch daemon, using the schema documented in
27 ovs-vswitchd.conf.db(5). The Open Virtual Network (OVN) project uses
28 two OVSDB schemas, documented as part of that project. Finally, Open
29 vSwitch includes the “VTEP” schema, documented in vtep(5) that many
30 third-party hardware switches support for configuring VXLAN, although
31 OVS itself does not directly use this schema.
32 The OVSDB protocol specification allows independent, interoperable
34 implementations of OVSDB to be developed. Open vSwitch includes an
35 OVSDB server implementation named ovsdb-server(1), which supports sev‐
36 eral protocol extensions documented in its manpage, and a basic com‐
37 mand-line OVSDB client named ovsdb-client(1), as well as OVSDB client
38 libraries for C and for Python. Open vSwitch documentation often
39 speaks of these OVSDB implementations in Open vSwitch as simply
40 “OVSDB,” even though that is distinct from the OVSDB protocol; we make
41 the distinction explicit only when it might otherwise be unclear from
42 the context.
43 In addition to these generic OVSDB server and client tools, Open
45 vSwitch includes tools for working with databases that have specific
46 schemas: ovs-vsctl works with the ovs-vswitchd configuration database
47 and vtep-ctl works with the VTEP database.
48 RFC 7047 specifies the OVSDB protocol but it does not specify an
50 on-disk storage format. Open vSwitch includes ovsdb-tool(1) for work‐
51 ing with its own on-disk database formats. The most notable feature of
52 this format is that ovsdb-tool(1) makes it easy for users to print the
53 transactions that have changed a database since the last time it was
54 compacted. This feature is often useful for troubleshooting.
55
57 Schemas in OVSDB have a JSON format that is specified in RFC 7047.
58 They are often stored in files with an extension .ovsschema. An
59 on-disk database in OVSDB includes a schema and data, embedding both
60 into a single file. The Open vSwitch utility ovsdb-tool has commands
61 that work with schema files and with the schemas embedded in database
62 files.
63 An Open vSwitch schema has three important identifiers. The first is
65 its name, which is also the name used in JSON-RPC calls to identify a
66 database based on that schema. For example, the schema used to config‐
67 ure Open vSwitch has the name Open_vSwitch. Schema names begin with a
68 letter or an underscore, followed by any number of letters, under‐
69 scores, or digits. The ovsdb-tool commands schema-name and db-name
70 extract the schema name from a schema or database file, respectively.
71 An OVSDB schema also has a version of the form x.y.z e.g. 1.2.3.
73 Schemas managed within the Open vSwitch project manage version number‐
74 ing in the following way (but OVSDB does not mandate this approach).
75 Whenever we change the database schema in a non-backward compatible way
76 (e.g. when we delete a column or a table), we increment <x> and set <y>
77 and <z> to 0. When we change the database schema in a backward compat‐
78 ible way (e.g. when we add a new column), we increment <y> and set <z>
79 to 0. When we change the database schema cosmetically (e.g. we rein‐
80 dent its syntax), we increment <z>. The ovsdb-tool commands
81 schema-version and db-version extract the schema version from a schema
82 or database file, respectively.
83 Very old OVSDB schemas do not have a version, but RFC 7047 mandates it.
85 An OVSDB schema optionally has a “checksum.” RFC 7047 does not specify
87 the use of the checksum and recommends that clients ignore it. Open
88 vSwitch uses the checksum to remind developers to update the version:
89 at build time, if the schema’s embedded checksum, ignoring the checksum
90 field itself, does not match the schema’s content, then it fails the
91 build with a recommendation to update the version and the checksum.
92 Thus, a developer who changes the schema, but does not update the ver‐
93 sion, receives an automatic reminder. In practice this has been an
94 effective way to ensure compliance with the version number policy. The
95 ovsdb-tool commands schema-cksum and db-cksum extract the schema check‐
96 sum from a schema or database file, respectively.
97
99 OVSDB supports three service models for databases: standalone,
100 active-backup, and clustered. The service models provide different
101 compromises among consistency, availability, and partition tolerance.
102 They also differ in the number of servers required and in terms of per‐
103 formance. The standalone and active-backup database service models
104 share one on-disk format, and clustered databases use a different for‐
105 mat, but the OVSDB programs work with both formats. ovsdb(5) documents
106 these file formats.
107 RFC 7047, which specifies the OVSDB protocol, does not mandate or spec‐
109 ify any particular service model.
110 The following sections describe the individual service models.
112 Standalone Database Service Model
114 A standalone database runs a single server. If the server stops run‐
115 ning, the database becomes inaccessible, and if the server’s storage is
116 lost or corrupted, the database’s content is lost. This service model
117 is appropriate when the database controls a process or activity to
118 which it is linked via “fate-sharing.” For example, an OVSDB instance
119 that controls an Open vSwitch virtual switch daemon, ovs-vswitchd, is a
120 standalone database because a server failure would take out both the
121 database and the virtual switch.
122 To set up a standalone database, use ovsdb-tool create to create a
124 database file, then run ovsdb-server to start the database service.
125 To configure a client, such as ovs-vswitchd or ovs-vsctl, to use a
127 standalone database, configure the server to listen on a “connection
128 method” that the client can reach, then point the client to that con‐
129 nection method. See Connection Methods below for information about
130 connection methods.
131 Active-Backup Database Service Model
133 An active-backup database runs two servers (on different hosts). At
134 any given time, one of the servers is designated with the active role
135 and the other the backup role. An active server behaves just like a
136 standalone server. A backup server makes an OVSDB connection to the
137 active server and uses it to continuously replicate its content as it
138 changes in real time. OVSDB clients can connect to either server but
139 only the active server allows data modification or lock transactions.
140 Setup for an active-backup database starts from a working standalone
142 database service, which is initially the active server. On another
143 node, to set up a backup server, create a database file with the same
144 schema as the active server. The initial contents of the database file
145 do not matter, as long as the schema is correct, so ovsdb-tool create
146 will work, as will copying the database file from the active server.
147 Then use ovsdb-server --sync-from=<active> to start the backup server,
148 where <active> is an OVSDB connection method (see Connection Methods
149 below) that connects to the active server. At that point, the backup
150 server will fetch a copy of the active database and keep it up-to-date
151 until it is killed.
152 When the active server in an active-backup server pair fails, an admin‐
154 istrator can switch the backup server to an active role with the
155 ovs-appctl command ovsdb-server/disconnect-active-ovsdb-server.
156 Clients then have read/write access to the now-active server. Of
157 course, administrators are slow to respond compared to software, so in
158 practice external management software detects the active server’s fail‐
159 ure and changes the backup server’s role. For example, the “Integra‐
160 tion Guide for Centralized Control” in the OVN documentation describes
161 how to use Pacemaker for this purpose in OVN.
162 Suppose an active server fails and its backup is promoted to active.
164 If the failed server is revived, it must be started as a backup server.
165 Otherwise, if both servers are active, then they may start out of sync,
166 if the database changed while the server was down, and they will con‐
167 tinue to diverge over time. This also happens if the software managing
168 the database servers cannot reach the active server and therefore
169 switches the backup to active, but other hosts can reach both servers.
170 These “split-brain” problems are unsolvable in general for server
171 pairs.
172 Compared to a standalone server, the active-backup service model some‐
174 what increases availability, at a risk of split-brain. It adds gener‐
175 ally insignificant performance overhead. On the other hand, the clus‐
176 tered service model, discussed below, requires at least 3 servers and
177 has greater performance overhead, but it avoids the need for external
178 management software and eliminates the possibility of split-brain.
179 Open vSwitch 2.6 introduced support for the active-backup service
181 model.
182 IMPORTANT:
184 There was a change of a database file format in version 2.15. To
185 upgrade/downgrade the ovsdb-server processes across this version
186 follow the instructions described under Upgrading from version 2.14
187 and earlier to 2.15 and later and Downgrading from version 2.15 and
188 later to 2.14 and earlier.
189 Clustered Database Service Model
191 A clustered database runs across 3 or 5 or more database servers (the
192 cluster) on different hosts. Servers in a cluster automatically syn‐
193 chronize writes within the cluster. A 3-server cluster can remain
194 available in the face of at most 1 server failure; a 5-server cluster
195 tolerates up to 2 failures. Clusters larger than 5 servers will also
196 work, with every 2 added servers allowing the cluster to tolerate 1
197 more failure, but write performance decreases. The number of servers
198 should be odd: a 4- or 6-server cluster cannot tolerate more failures
199 than a 3- or 5-server cluster, respectively.
200 To set up a clustered database, first initialize it on a single node by
202 running ovsdb-tool create-cluster, then start ovsdb-server. Depending
203 on its arguments, the create-cluster command can create an empty data‐
204 base or copy a standalone database’s contents into the new database.
205 To configure a client to use a clustered database, first configure all
207 of the servers to listen on a connection method that the client can
208 reach, then point the client to all of the servers’ connection methods,
209 comma-separated. See Connection Methods, below, for more detail.
210 Open vSwitch 2.9 introduced support for the clustered service model.
212 How to Maintain a Clustered Database
214 To add a server to a cluster, run ovsdb-tool join-cluster on the new
215 server and start ovsdb-server. To remove a running server from a clus‐
216 ter, use ovs-appctl to invoke the cluster/leave command. When a server
217 fails and cannot be recovered, e.g. because its hard disk crashed, or
218 to otherwise remove a server that is down from a cluster, use
219 ovs-appctl to invoke cluster/kick to make the remaining servers kick it
220 out of the cluster.
221 The above methods for adding and removing servers only work for healthy
223 clusters, that is, for clusters with no more failures than their maxi‐
224 mum tolerance. For example, in a 3-server cluster, the failure of 2
225 servers prevents servers joining or leaving the cluster (as well as
226 database access). To prevent data loss or inconsistency, the preferred
227 solution to this problem is to bring up enough of the failed servers to
228 make the cluster healthy again, then if necessary remove any remaining
229 failed servers and add new ones. If this cannot be done, though, use
230 ovs-appctl to invoke cluster/leave --force on a running server. This
231 command forces the server to which it is directed to leave its cluster
232 and form a new single-node cluster that contains only itself. The data
233 in the new cluster may be inconsistent with the former cluster: trans‐
234 actions not yet replicated to the server will be lost, and transactions
235 not yet applied to the cluster may be committed. Afterward, any
236 servers in its former cluster will regard the server to have failed.
237 Once a server leaves a cluster, it may never rejoin it. Instead, cre‐
239 ate a new server and join it to the cluster.
240 The servers in a cluster synchronize data over a cluster management
242 protocol that is specific to Open vSwitch; it is not the same as the
243 OVSDB protocol specified in RFC 7047. For this purpose, a server in a
244 cluster is tied to a particular IP address and TCP port, which is spec‐
245 ified in the ovsdb-tool command that creates or joins the cluster. The
246 TCP port used for clustering must be different from that used for OVSDB
247 clients. To change the port or address of a server in a cluster, first
248 remove it from the cluster, then add it back with the new address.
249 To upgrade the ovsdb-server processes in a cluster from one version of
251 Open vSwitch to another, upgrading them one at a time will keep the
252 cluster healthy during the upgrade process. (This is different from
253 upgrading a database schema, which is covered later under Upgrading or
254 Downgrading a Database.)
255 IMPORTANT:
257 There was a change of a database file format in version 2.15. To
258 upgrade/downgrade the ovsdb-server processes across this version
259 follow the instructions described under Upgrading from version 2.14
260 and earlier to 2.15 and later and Downgrading from version 2.15 and
261 later to 2.14 and earlier.
262 Clustered OVSDB does not support the OVSDB “ephemeral columns” feature.
264 ovsdb-tool and ovsdb-client change ephemeral columns into persistent
265 ones when they work with schemas for clustered databases. Future ver‐
266 sions of OVSDB might add support for this feature.
267 Upgrading from version 2.14 and earlier to 2.15 and later
269 There is a change of a database file format in version 2.15 that
270 doesn’t allow older versions of ovsdb-server to read the database file
271 modified by the ovsdb-server version 2.15 or later. This also affects
272 runtime communications between servers in active-backup and cluster
273 service models. To upgrade the ovsdb-server processes from one version
274 of Open vSwitch (2.14 or earlier) to another (2.15 or higher) instruc‐
275 tions below should be followed. (This is different from upgrading a
276 database schema, which is covered later under Upgrading or Downgrading
277 a Database.)
278 In case of standalone service model no special handling during upgrade
280 is required.
281 For the active-backup service model, administrator needs to update
283 backup ovsdb-server first and the active one after that, or shut down
284 both servers and upgrade at the same time.
285 For the cluster service model recommended upgrade strategy is follow‐
287 ing:
288 1. Upgrade processes one at a time. Each ovsdb-server process after
290 upgrade should be started with --disable-file-column-diff command
291 line argument.
292 2. When all ovsdb-server processes upgraded, use ovs-appctl to invoke
294 ovsdb/file/column-diff-enable command on each of them or restart all
295 ovsdb-server processes one at a time without --disable-file-col‐
296 umn-diff command line option.
297 Downgrading from version 2.15 and later to 2.14 and earlier
299 Similar to upgrading covered under Upgrading from version 2.14 and ear‐
300 lier to 2.15 and later, downgrading from the ovsdb-server version 2.15
301 and later to 2.14 and earlier requires additional steps. (This is dif‐
302 ferent from upgrading a database schema, which is covered later under
303 Upgrading or Downgrading a Database.)
304 For all service models it’s required to:
306 1. Stop all ovsdb-server processes (single process for standalone ser‐
308 vice model, all involved processes for active-backup and cluster
309 service models).
310 2. Compact all database files with ovsdb-tool compact command.
312 3. Downgrade and restart ovsdb-server processes.
314 Understanding Cluster Consistency
316 To ensure consistency, clustered OVSDB uses the Raft algorithm
317 described in Diego Ongaro’s Ph.D. thesis, “Consensus: Bridging Theory
318 and Practice”. In an operational Raft cluster, at any given time a
319 single server is the “leader” and the other nodes are “followers”.
320 Only the leader processes transactions, but a transaction is only com‐
321 mitted when a majority of the servers confirm to the leader that they
322 have written it to persistent storage.
323 In most database systems, read and write access to the database happens
325 through transactions. In such a system, Raft allows a cluster to
326 present a strongly consistent transactional interface. OVSDB uses con‐
327 ventional transactions for writes, but clients often effectively do
328 reads a different way, by asking the server to “monitor” a database or
329 a subset of one on the client’s behalf. Whenever monitored data
330 changes, the server automatically tells the client what changed, which
331 allows the client to maintain an accurate snapshot of the database in
332 its memory. Of course, at any given time, the snapshot may be somewhat
333 dated since some of it could have changed without the change notifica‐
334 tion yet being received and processed by the client.
335 Given this unconventional usage model, OVSDB also adopts an unconven‐
337 tional clustering model. Each server in a cluster acts independently
338 for the purpose of monitors and read-only transactions, without verify‐
339 ing that data is up-to-date with the leader. Servers forward transac‐
340 tions that write to the database to the leader for execution, ensuring
341 consistency. This has the following consequences:
342 · Transactions that involve writes, against any server in the cluster,
344 are linearizable if clients take care to use correct prerequisites,
345 which is the same condition required for linearizability in a stand‐
346 alone OVSDB. (Actually, “at-least-once” consistency, because OVSDB
347 does not have a session mechanism to drop duplicate transactions if a
348 connection drops after the server commits it but before the client
349 receives the result.)
350 · Read-only transactions can yield results based on a stale version of
352 the database, if they are executed against a follower. Transactions
353 on the leader always yield fresh results. (With monitors, as
354 explained above, a client can always see stale data even without
355 clustering, so clustering does not change the consistency model for
356 monitors.)
357 · Monitor-based (or read-heavy) workloads scale well across a cluster,
359 because clustering OVSDB adds no additional work or communication for
360 reads and monitors.
361 · A write-heavy client should connect to the leader, to avoid the over‐
363 head of followers forwarding transactions to the leader.
364 · When a client conducts a mix of read and write transactions across
366 more than one server in a cluster, it can see inconsistent results
367 because a read transaction might read stale data whose updates have
368 not yet propagated from the leader. By default, utilities such as
369 ovn-sbctl (in OVN) connect to the cluster leader to avoid this issue.
370 The same might occur for transactions against a single follower
372 except that the OVSDB server ensures that the results of a write for‐
373 warded to the leader by a given server are visible at that server
374 before it replies to the requesting client.
375 · If a client uses a database on one server in a cluster, then another
377 server in the cluster (perhaps because the first server failed), the
378 client could observe stale data. Clustered OVSDB clients, however,
379 can use a column in the _Server database to detect that data on a
380 server is older than data that the client previously read. The OVSDB
381 client library in Open vSwitch uses this feature to avoid servers
382 with stale data.
383
385 OVSDB can layer replication on top of any of its service models.
386 Replication, in this context, means to make, and keep up-to-date, a
387 read-only copy of the contents of a database (the replica). One use of
388 replication is to keep an up-to-date backup of a database. A replica
389 used solely for backup would not need to support clients of its own. A
390 set of replicas that do serve clients could be used to scale out read
391 access to the primary database.
392 A database replica is set up in the same way as a backup server in an
394 active-backup pair, with the difference that the replica is never pro‐
395 moted to an active role.
396 A database can have multiple replicas.
398 Open vSwitch 2.6 introduced support for database replication.
400
402 An OVSDB connection method is a string that specifies how to make a
403 JSON-RPC connection between an OVSDB client and server. Connection
404 methods are part of the Open vSwitch implementation of OVSDB and not
405 specified by RFC 7047. ovsdb-server uses connection methods to specify
406 how it should listen for connections from clients and ovsdb-client uses
407 them to specify how it should connect to a server. Connections in the
408 opposite direction, where ovsdb-server connects to a client that is
409 configured to listen for an incoming connection, are also possible.
410 Connection methods are classified as active or passive. An active con‐
412 nection method makes an outgoing connection to a remote host; a passive
413 connection method listens for connections from remote hosts. The most
414 common arrangement is to configure an OVSDB server with passive connec‐
415 tion methods and clients with active ones, but the OVSDB implementation
416 in Open vSwitch supports the opposite arrangement as well.
417 OVSDB supports the following active connection methods:
419 ssl:<host>:<port>
421 The specified SSL or TLS <port> on the given <host>.
422 tcp:<host>:<port>
424 The specified TCP <port> on the given <host>.
425 unix:<file>
427 On Unix-like systems, connect to the Unix domain server socket
428 named <file>.
429 On Windows, connect to a local named pipe that is represented by
431 a file created in the path <file> to mimic the behavior of a
432 Unix domain socket.
433 <method1>,<method2>,…,<methodN>
435 For a clustered database service to be highly available, a
436 client must be able to connect to any of the servers in the
437 cluster. To do so, specify connection methods for each of the
438 servers separated by commas (and optional spaces).
439 In theory, if machines go up and down and IP addresses change in
441 the right way, a client could talk to the wrong instance of a
442 database. To avoid this possibility, add cid:<uuid> to the list
443 of methods, where <uuid> is the cluster ID of the desired data‐
444 base cluster, as printed by ovsdb-tool db-cid. This feature is
445 optional.
446 OVSDB supports the following passive connection methods:
448 pssl:<port>[:<ip>]
450 Listen on the given TCP <port> for SSL or TLS connections. By
451 default, connections are not bound to a particular local IP
452 address. Specifying <ip> limits connections to those from the
453 given IP.
454 ptcp:<port>[:<ip>]
456 Listen on the given TCP <port>. By default, connections are not
457 bound to a particular local IP address. Specifying <ip> limits
458 connections to those from the given IP.
459 punix:<file>
461 On Unix-like systems, listens for connections on the Unix domain
462 socket named <file>.
463 On Windows, listens on a local named pipe, creating a named pipe
465 <file> to mimic the behavior of a Unix domain socket. The ACLs
466 of the named pipe include LocalSystem, Administrators, and Cre‐
467 ator Owner.
468 All IP-based connection methods accept IPv4 and IPv6 addresses. To
470 specify an IPv6 address, wrap it in square brackets, e.g.
471 ssl:[::1]:6640. Passive IP-based connection methods by default listen
472 for IPv4 connections only; use [::] as the address to accept both IPv4
473 and IPv6 connections, e.g. pssl:6640:[::]. DNS names are also accepted
474 if built with unbound library. On Linux, use %<device> to designate a
475 scope for IPv6 link-level addresses, e.g. ssl:[fe80::1234%eth0]:6653.
476 The <port> may be omitted from connection methods that use a port num‐
478 ber. The default <port> for TCP-based connection methods is 6640, e.g.
479 pssl: is equivalent to pssl:6640. In Open vSwitch prior to version
480 2.4.0, the default port was 6632. To avoid incompatibility between
481 older and newer versions, we encourage users to specify a port number.
482 The ssl and pssl connection methods requires additional configuration
484 through --private-key, --certificate, and --ca-cert command line
485 options. Open vSwitch can be built without SSL support, in which case
486 these connection methods are not supported.
487
489 This section describes how to handle various events in the life cycle
490 of a database using the Open vSwitch implementation of OVSDB.
491 Creating a Database
493 Creating and starting up the service for a new database was covered
494 separately for each database service model in the Service Models sec‐
495 tion, above.
496 Backing Up and Restoring a Database
498 OVSDB is often used in contexts where the database contents are not
499 particularly valuable. For example, in many systems, the database for
500 configuring ovs-vswitchd is essentially rebuilt from scratch at boot
501 time. It is not worthwhile to back up these databases.
502 When OVSDB is used for valuable data, a backup strategy is worth con‐
504 sidering. One way is to use database replication, discussed above in
505 Database Replication which keeps an online, up-to-date copy of a data‐
506 base, possibly on a remote system. This works with all OVSDB service
507 models.
508 A more common backup strategy is to periodically take and store a snap‐
510 shot. For the standalone and active-backup service models, making a
511 copy of the database file, e.g. using cp, effectively makes a snapshot,
512 and because OVSDB database files are append-only, it works even if the
513 database is being modified when the snapshot takes place. This
514 approach does not work for clustered databases.
515 Another way to make a backup, which works with all OVSDB service mod‐
517 els, is to use ovsdb-client backup, which connects to a running data‐
518 base server and outputs an atomic snapshot of its schema and content,
519 in the same format used for standalone and active-backup databases.
520 Multiple options are also available when the time comes to restore a
522 database from a backup. For the standalone and active-backup service
523 models, one option is to stop the database server or servers, overwrite
524 the database file with the backup (e.g. with cp), and then restart the
525 servers. Another way, which works with any service model, is to use
526 ovsdb-client restore, which connects to a running database server and
527 replaces the data in one of its databases by a provided snapshot. The
528 advantage of ovsdb-client restore is that it causes zero downtime for
529 the database and its server. It has the downside that UUIDs of rows in
530 the restored database will differ from those in the snapshot, because
531 the OVSDB protocol does not allow clients to specify row UUIDs.
532 None of these approaches saves and restores data in columns that the
534 schema designates as ephemeral. This is by design: the designer of a
535 schema only marks a column as ephemeral if it is acceptable for its
536 data to be lost when a database server restarts.
537 Clustering and backup serve different purposes. Clustering increases
539 availability, but it does not protect against data loss if, for exam‐
540 ple, a malicious or malfunctioning OVSDB client deletes or tampers with
541 data.
542 Changing Database Service Model
544 Use ovsdb-tool create-cluster to create a clustered database from the
545 contents of a standalone database. Use ovsdb-client backup to create a
546 standalone database from the contents of a running clustered database.
547 When the cluster is down and cannot be revived, ovsdb-client backup
548 will not work.
549 Use ovsdb-tool cluster-to-standalone to convert clustered database to
551 standalone database when the cluster is down and cannot be revived.
552 Upgrading or Downgrading a Database
554 The evolution of a piece of software can require changes to the schemas
555 of the databases that it uses. For example, new features might require
556 new tables or new columns in existing tables, or conceptual changes
557 might require a database to be reorganized in other ways. In some
558 cases, the easiest way to deal with a change in a database schema is to
559 delete the existing database and start fresh with the new schema, espe‐
560 cially if the data in the database is easy to reconstruct. But in many
561 other cases, it is better to convert the database from one schema to
562 another.
563 The OVSDB implementation in Open vSwitch has built-in support for some
565 simple cases of converting a database from one schema to another. This
566 support can handle changes that add or remove database columns or
567 tables or that eliminate constraints (for example, changing a column
568 that must have exactly one value into one that has one or more values).
569 It can also handle changes that add constraints or make them stricter,
570 but only if the existing data in the database satisfies the new con‐
571 straints (for example, changing a column that has one or more values
572 into a column with exactly one value, if every row in the column has
573 exactly one value). The built-in conversion can cause data loss in
574 obvious ways, for example if the new schema removes tables or columns,
575 or indirectly, for example by deleting unreferenced rows in tables that
576 the new schema marks for garbage collection.
577 Converting a database can lose data, so it is wise to make a backup
579 beforehand.
580 To use OVSDB’s built-in support for schema conversion with a standalone
582 or active-backup database, first stop the database server or servers,
583 then use ovsdb-tool convert to convert it to the new schema, and then
584 restart the database server.
585 OVSDB also supports online database schema conversion for any of its
587 database service models. To convert a database online, use
588 ovsdb-client convert. The conversion is atomic, consistent, isolated,
589 and durable. ovsdb-server disconnects any clients connected when the
590 conversion takes place (except clients that use the set_db_change_aware
591 Open vSwitch extension RPC). Upon reconnection, clients will discover
592 that the schema has changed.
593 Schema versions and checksums (see Schemas above) can give hints about
595 whether a database needs to be converted to a new schema. If there is
596 any question, though, the needs-conversion command on ovsdb-tool and
597 ovsdb-client can provide a definitive answer.
598 Working with Database History
600 Both on-disk database formats that OVSDB supports are organized as a
601 stream of transaction records. Each record describes a change to the
602 database as a list of rows that were inserted or deleted or modified,
603 along with the details. Therefore, in normal operation, a database
604 file only grows, as each change causes another record to be appended at
605 the end. Usually, a user has no need to understand this file struc‐
606 ture. This section covers some exceptions.
607 Compacting Databases
609 If OVSDB database files were truly append-only, then over time they
610 would grow without bound. To avoid this problem, OVSDB can compact a
611 database file, that is, replace it by a new version that contains only
612 the current database contents, as if it had been inserted by a single
613 transaction. From time to time, ovsdb-server automatically compacts a
614 database that grows much larger than its minimum size.
615 Because ovsdb-server automatically compacts databases, it is usually
617 not necessary to compact them manually, but OVSDB still offers a few
618 ways to do it. First, ovsdb-tool compact can compact a standalone or
619 active-backup database that is not currently being served by
620 ovsdb-server (or otherwise locked for writing by another process). To
621 compact any database that is currently being served by ovsdb-server,
622 use ovs-appctl to send the ovsdb-server/compact command. Each server
623 in an active-backup or clustered database maintains its database file
624 independently, so to compact all of them, issue this command separately
625 on each server.
626 Viewing History
628 The ovsdb-tool utility’s show-log command displays the transaction
629 records in an OVSDB database file in a human-readable format. By
630 default, it shows minimal detail, but adding the option -m once or
631 twice increases the level of detail. In addition to the transaction
632 data, it shows the time and date of each transaction and any “comment”
633 added to the transaction by the client. The comments can be helpful
634 for quickly understanding a transaction; for example, ovs-vsctl adds
635 its command line to the transactions that it makes.
636 The show-log command works with both OVSDB file formats, but the
638 details of the output format differ. For active-backup and clustered
639 databases, the sequence of transactions in each server’s log will dif‐
640 fer, even at points when they reflect the same data.
641 Truncating History
643 It may occasionally be useful to “roll back” a database file to an ear‐
644 lier point. Because of the organization of OVSDB records, this is easy
645 to do. Start by noting the record number <i> of the first record to
646 delete in ovsdb-tool show-log output. Each record is two lines of
647 plain text, so trimming the log is as simple as running head -n <j>,
648 where <j> = 2 * <i>.
649 Corruption
651 When ovsdb-server opens an OVSDB database file, of any kind, it reads
652 as many transaction records as it can from the file until it reaches
653 the end of the file or it encounters a corrupted record. At that point
654 it stops reading and regards the data that it has read to this point as
655 the full contents of the database file, effectively rolling the data‐
656 base back to an earlier point.
657 Each transaction record contains an embedded SHA-1 checksum, which the
659 server verifies as it reads a database file. It detects corruption
660 when a checksum fails to verify. Even though SHA-1 is no longer con‐
661 sidered secure for use in cryptography, it is acceptable for this pur‐
662 pose because it is not used to defend against malicious attackers.
663 The first record in a standalone or active-backup database file speci‐
665 fies the schema. ovsdb-server will refuse to work with a database
666 where this record is corrupted, or with a clustered database file with
667 corruption in the first few records. Delete and recreate such a data‐
668 base, or restore it from a backup.
669 When ovsdb-server adds records to a database file in which it detected
671 corruption, it first truncates the file just after the last good
672 record.
673
675 RFC 7047, “The Open vSwitch Database Management Protocol.”
676 Open vSwitch implementations of generic OVSDB functionality:
678 ovsdb-server(1), ovsdb-client(1), ovsdb-tool(1).
679 Tools for working with databases that have specific OVSDB schemas:
681 ovs-vsctl(8), vtep-ctl(8), and (in OVN) ovn-nbctl(8), ovn-sbctl(8).
682 OVSDB schemas for Open vSwitch and related functionality:
684 ovs-vswitchd.conf.db(5), vtep(5), and (in OVN) ovn-nb(5), ovn-sb(5).
685
687 The Open vSwitch Development Community
688
690 2021, The Open vSwitch Development Community
6912.15 Feb 17, 2021 OVSDB(7)