4 pg_resetwal — reset the write-ahead log and other control information
5 of a PostgreSQL database cluster
9 pg_resetwal [ -f | --force ] [ -n | --dry-run ] [option...] [ -D |
14 pg_resetwal clears the write-ahead log (WAL) and optionally resets some
15 other control information stored in the pg_control file. This function
16 is sometimes needed if these files have become corrupted. It should be
17 used only as a last resort, when the server will not start due to such
20 Some options, such as --wal-segsize (see below), can also be used to
21 modify certain global settings of a database cluster without the need
22 to rerun initdb. This can be done safely on an otherwise sound database
23 cluster, if none of the dangerous modes mentioned below are used.
25 If pg_resetwal is used on a data directory where the server has been
26 cleanly shut down and the control file is sound, then it will have no
27 effect on the contents of the database system, except that no longer
28 used WAL files are cleared away. Any other use is potentially dangerous
29 and must be done with great care. pg_resetwal will require the -f
30 (force) option to be specified before working on a data directory in an
31 unclean shutdown state or with a corrupted control file.
33 After running this command on a data directory with corrupted WAL or a
34 corrupted control file, it should be possible to start the server, but
35 bear in mind that the database might contain inconsistent data due to
36 partially-committed transactions. You should immediately dump your
37 data, run initdb, and restore. After restore, check for inconsistencies
40 If pg_resetwal complains that it cannot determine valid data for
41 pg_control, you can force it to proceed anyway by specifying the -f
42 (force) option. In this case plausible values will be substituted for
43 the missing data. Most of the fields can be expected to match, but
44 manual assistance might be needed for the next OID, next transaction ID
45 and epoch, next multitransaction ID and offset, and WAL starting
46 location fields. These fields can be set using the options discussed
47 below. If you are not able to determine correct values for all these
48 fields, -f can still be used, but the recovered database must be
49 treated with even more suspicion than usual: an immediate dump and
50 restore is imperative. Do not execute any data-modifying operations in
51 the database before you dump, as any such action is likely to make the
54 This utility can only be run by the user who installed the server,
55 because it requires read/write access to the data directory.
62 Specifies the location of the database directory. For safety
63 reasons, you must specify the data directory on the command
64 line. pg_resetwal does not use the environment variable PGDATA.
68 Force pg_resetwal to proceed even in situations where it could
69 be dangerous, as explained above. Specifically, this option is
70 required to proceed if the server had not been cleanly shut down
71 or if pg_resetwal cannot determine valid data for pg_control.
75 The -n/--dry-run option instructs pg_resetwal to print the
76 values reconstructed from pg_control and values about to be
77 changed, and then exit without modifying anything. This is
78 mainly a debugging tool, but can be useful as a sanity check
79 before allowing pg_resetwal to proceed for real.
83 Display version information, then exit.
89 The following options are only needed when pg_resetwal is unable to
90 determine appropriate values by reading pg_control. Safe values can be
91 determined as described below. For values that take numeric arguments,
92 hexadecimal values can be specified by using the prefix 0x. Note that
93 these instructions only apply with the standard block size of 8 kB.
96 --commit-timestamp-ids=xid,xid
97 Manually set the oldest and newest transaction IDs for which the
98 commit time can be retrieved.
100 A safe value for the oldest transaction ID for which the commit
101 time can be retrieved (first part) can be determined by looking
102 for the numerically smallest file name in the directory
103 pg_commit_ts under the data directory. Conversely, a safe value
104 for the newest transaction ID for which the commit time can be
105 retrieved (second part) can be determined by looking for the
106 numerically greatest file name in the same directory. The file
107 names are in hexadecimal.
111 Manually set the next transaction ID's epoch.
113 The transaction ID epoch is not actually stored anywhere in the
114 database except in the field that is set by pg_resetwal, so any
115 value will work so far as the database itself is concerned. You
116 might need to adjust this value to ensure that replication
117 systems such as Slony-I and Skytools work correctly — if so, an
118 appropriate value should be obtainable from the state of the
119 downstream replicated database.
122 --next-wal-file=walfile
123 Manually set the WAL starting location by specifying the name of
124 the next WAL segment file.
126 The name of next WAL segment file should be larger than any WAL
127 segment file name currently existing in the directory pg_wal
128 under the data directory. These names are also in hexadecimal
129 and have three parts. The first part is the “timeline ID” and
130 should usually be kept the same. For example, if
131 00000001000000320000004A is the largest entry in pg_wal, use -l
132 00000001000000320000004B or higher.
134 Note that when using nondefault WAL segment sizes, the numbers
135 in the WAL file names are different from the LSNs that are
136 reported by system functions and system views. This option takes
137 a WAL file name, not an LSN.
141 pg_resetwal itself looks at the files in pg_wal and chooses a
142 default -l setting beyond the last existing file name.
143 Therefore, manual adjustment of -l should only be needed if you
144 are aware of WAL segment files that are not currently present in
145 pg_wal, such as entries in an offline archive; or if the
146 contents of pg_wal have been lost entirely.
149 --multixact-ids=mxid,mxid
150 Manually set the next and oldest multitransaction ID.
152 A safe value for the next multitransaction ID (first part) can
153 be determined by looking for the numerically largest file name
154 in the directory pg_multixact/offsets under the data directory,
155 adding one, and then multiplying by 65536 (0x10000). Conversely,
156 a safe value for the oldest multitransaction ID (second part of
157 -m) can be determined by looking for the numerically smallest
158 file name in the same directory and multiplying by 65536. The
159 file names are in hexadecimal, so the easiest way to do this is
160 to specify the option value in hexadecimal and append four
165 Manually set the next OID.
167 There is no comparably easy way to determine a next OID that's
168 beyond the largest one in the database, but fortunately it is
169 not critical to get the next-OID setting right.
172 --multixact-offset=mxoff
173 Manually set the next multitransaction offset.
175 A safe value can be determined by looking for the numerically
176 largest file name in the directory pg_multixact/members under
177 the data directory, adding one, and then multiplying by 52352
178 (0xCC80). The file names are in hexadecimal. There is no simple
179 recipe such as the ones for other options of appending zeroes.
182 --oldest-transaction-id=xid
183 Manually set the oldest unfrozen transaction ID.
185 A safe value can be determined by looking for the numerically
186 smallest file name in the directory pg_xact under the data
187 directory and then multiplying by 1048576 (0x100000). Note that
188 the file names are in hexadecimal. It is usually easiest to
189 specify the option value in hexadecimal too. For example, if
190 0007 is the smallest entry in pg_xact, -u 0x700000 will work
191 (five trailing zeroes provide the proper multiplier).
194 --next-transaction-id=xid
195 Manually set the next transaction ID.
197 A safe value can be determined by looking for the numerically
198 largest file name in the directory pg_xact under the data
199 directory, adding one, and then multiplying by 1048576
200 (0x100000). Note that the file names are in hexadecimal. It is
201 usually easiest to specify the option value in hexadecimal too.
202 For example, if 0011 is the largest entry in pg_xact, -x
203 0x1200000 will work (five trailing zeroes provide the proper
206 --char-signedness=option
207 Manually set the default char signedness. Possible values are
210 For a database cluster that pg_upgrade upgraded from a
211 PostgreSQL version before 18, the safe value would be the
212 default char signedness of the platform that ran the cluster
213 before that upgrade. For all other clusters, signed would be the
214 safe value. However, this option is exclusively for use with
215 pg_upgrade and should not normally be used manually.
217 --wal-segsize=wal_segment_size
218 Set the new WAL segment size, in megabytes. The value must be
219 set to a power of 2 between 1 and 1024 (megabytes). See the same
220 option of initdb for more information.
222 This option can also be used to change the WAL segment size of
223 an existing database cluster, avoiding the need to re-initdb.
227 While pg_resetwal will set the WAL starting address beyond the
228 latest existing WAL segment file, some segment size changes can
229 cause previous WAL file names to be reused. It is recommended to
230 use -l together with this option to manually set the WAL
231 starting address if WAL file name overlap will cause problems
232 with your archiving strategy.
237 Specifies whether to use color in diagnostic messages. Possible
238 values are always, auto and never.
242 This command must not be used when the server is running. pg_resetwal
243 will refuse to start up if it finds a server lock file in the data
244 directory. If the server crashed then a lock file might have been left
245 behind; in that case you can remove the lock file to allow pg_resetwal
246 to run. But before you do so, make doubly certain that there is no
247 server process still alive.
249 pg_resetwal works only with servers of the same major version.