3 salvageserver - Initializes the Salvageserver component of the dafs process
10 B<salvageserver> [I<initcmd>] S<<< [B<-partition> <I<name of partition to salvage>>] >>>
11 S<<< [B<-volumeid> <I<volume id to salvage>>] >>> [B<-debug>] [B<-nowrite>]
12 [B<-inodes>] [B<-force>] [B<-oktozap>] [B<-rootinodes>]
13 [B<-salvagedirs>] [B<-blockreads>]
14 S<<< [B<-parallel> <I<# of max parallel partition salvaging>>] >>>
15 S<<< [B<-tmpdir> <I<name of dir to place tmp files>>] >>>
16 [B<-showlog>] [B<-showsuid>] [B<-showmounts>]
17 S<<< [B<-orphans> (ignore | remove | attach)] >>>
19 S<<< [B<-syslogfacility> <I<Syslog facility number to use>>] >>>
20 [B<-client>] [B<-help>]
27 In its typical mode of operation, the B<salvageserver> is a daemon process
28 responsible for salvaging volumes. It is a component of the C<dafs>
29 process type. In the conventional configuration, its binary file is
30 located in the F</usr/afs/bin> directory on a file server machine.
32 The Salvageserver daemon is responsible for scheduling and executing
33 volume salvage operations on behalf of client processes. The fileserver
34 acts as the primary salvageserver client: any failed volume attach
35 operation results in a salvageserver scheduling request. The
36 salvageserver also accepts periodic volume activity messages in order to
37 update its salvage request priority queue. Other clients of the
38 salvageserver daemon include the B<salvsync-debug> utility, and the
39 salvageserver command itself by passing the B<-client> flag.
41 The salvage operations performed on vice partition data are nearly
42 identical to those performed by the standalone Salvager command. The
43 key differences between the two commands are:
49 The Salvageserver is a daemon process which runs concurrently with the
50 fileserver. In contrast, the Salvager is a stand-alone application which
51 is invoked when the fileserver and volserver are not running.
55 The Salvageserver is incapable of performing whole partition salvage
56 operations; it operates at volume group granularity.
60 The Salvageserver normally creates new inodes as it repairs damage. If the
61 partition is so full that there is no room for new inodes, use the
62 B<-nowrite> argument to bringing undamaged volumes online without
63 attempting to salvage damaged volumes. Then use the B<vos move> command to
64 move one or more of the undamaged volumes to other partitions, freeing up
65 the space that the Salvageserver needs to create new inodes.
67 By default, multiple Salvageserver subprocesses run in parallel: one for each
68 volume group. By default, four concurrent salvage operations are
69 permitted. You may alter this default by providing a positive integer
70 value for the B<-parallel> argument. The maximum permitted value is 32
71 concurrent salvageserver subprocesses.
73 By default, the salvageserver enables a heuristic which attempts to stop
74 disk head thrashing by concurrent salvageserver subprocesses. Unfortunately,
75 this heuristic significantly degrades performance in many cases. In at least
76 the following environments, passing the C<all> string to the B<-parallel>
77 argument is strongly encouraged:
87 When a vice partition is backed by multiple disks (e.g. RAID)
91 When a vice partition is backed by SAN-attached storage, LVM, or some other
92 form of storage virtualization which would cause unix device id numbers to
97 The Salvageserver creates temporary files as it runs, by default writing them
98 to the partition it is salvaging. The number of files can be quite large,
99 and if the partition is too full to accommodate them, the Salvageserver
100 terminates without completing the salvage operation (it always removes the
101 temporary files before exiting). Other Salvageserver subprocesses running at
102 the same time continue until they finish salvaging all other partitions
103 where there is enough disk space for temporary files. To complete the
104 interrupted salvage, reissue the command against the appropriate
105 partitions, adding the B<-tmpdir> argument to redirect the temporary files
106 to a local disk directory that has enough space.
108 The B<-orphans> argument controls how the Salvageserver handles orphaned files
109 and directories that it finds on server partitions it is salvaging. An
110 I<orphaned> element is completely inaccessible because it is not
111 referenced by the vnode of any directory that can act as its parent (is
112 higher in the filespace). Orphaned objects occupy space on the server
113 partition, but do not count against the volume's quota.
115 To generate a list of all mount points that reside in one or more volumes,
116 rather than actually salvaging them, include the B<-showmounts> flag.
118 This command does not use the syntax conventions of the AFS command
119 suites. Provide the command name and all option names in full.
127 Accommodates the command's use of the AFS command parser, and is optional.
129 =item B<-partition> <I<name of partition to salvage>>
131 Specifies the name of the partition to salvage. Specify the full partition
132 name using the form F</vicepI<x>> or F</vicepI<xx>>. Omit this argument to
133 salvage every partition on the file server machine.
135 =item B<-volumeid> <I<volume id to salvage>>
137 Specifies the volume ID of a specific read/write volume to salvage. The
138 B<-partition> argument must be provided along with this one and specify
139 the volume's actual site.
143 This flag should be considered deprecated. Its primary purpose was to disable
144 forking and parallelization of the Salvager so that log messages were not
145 interleaved. Due to the manner in which F</usr/afs/logs/SalsrvLog> is
146 written, log messages from subprocesses are never interleaved; the entire log
147 for a volume group salvage is appended to the master log as one atomic
152 Brings all undamaged volumes online without attempting to salvage any
157 Records in the F</usr/afs/logs/SalsrvLog> file a list of all AFS inodes
158 that the Salvageserver modified.
162 Inspects all volumes for corruption, not just those that are marked as
163 having been active when a crash occurred.
167 Removes a volume that is so damaged that even issuing the B<vos zap>
168 command with the B<-force> flag is ineffective. Combine it with the
169 B<-partition> and B<-volumeid> arguments to identify the volume to remove.
170 Using this flag will destroy data that cannot be read, so use only with
171 caution and when you're certain that nothing in that volume is still
176 Records in the F</usr/afs/logs/SalsrvLog> file a list of all AFS inodes
177 owned by the local superuser C<root>.
179 =item B<-salvagedirs>
181 Salvages entire directory structures, even if they do not appear to be
182 damaged. By default, the Salvageserver salvages a directory only if it is
183 flagged as corrupted.
187 Forces the Salvageserver to read a partition one disk block (512 bytes) at a
188 time and to skip any blocks that are too badly damaged to be salvaged.
189 This allows it to salvage as many volumes as possible. By default, the
190 Salvageserver reads large disk blocks, which can cause it to exit prematurely
191 if it encounters disk errors. Use this flag if the partition to be
192 salvaged has disk errors.
194 =item B<-parallel> <I<# of max parallel partition salvaging>>
196 Specifies the maximum number of Salvageserver subprocesses to run in parallel.
197 Provide one of three values:
203 An integer from the range C<1> to C<32>. A value of C<1> means that a
204 single Salvageserver subprocess salvages the volume groups sequentially.
205 The disk partition heuristic (see above) based upon unix device ids is
210 The disk partition heuristic (see above) based upon unix device ids is
215 The string C<all> followed immediately (with no intervening space) by an
216 integer from the range C<1> to C<32>, to run the specified number of
217 Salvageserver subprocesses in parallel on volume groups. The disk partition
218 heuristic (see above) based upon unix device ids is disabled.
222 If this argument is omitted, up to four Salvageserver subprocesses run
225 =item B<-tmpdir> <I<name of dir to place tmp files>>
227 Names a local disk directory in which the Salvageserver places the temporary
228 files it creates during a salvage operation, instead of writing them to
229 the partition being salvaged (the default). If the Salvageserver cannot write
230 to the specified directory, it attempts to write to the partition being
235 Displays on the standard output stream all log data that is being written
236 to the F</usr/afs/logs/SalsrvLog> file.
240 Displays a list of the pathnames for all files that have the setuid or
245 Records in the F</usr/afs/logs/SalsrvLog> file all mount points found in
246 each volume. The Salvageserver does not repair corruption in the volumes, if
249 =item B<-orphans> (ignore | remove | attach)
251 Controls how the Salvageserver handles orphaned files and directories. Choose
252 one of the following three values:
258 Leaves the orphaned objects on the disk, but prints a message to the
259 F</usr/afs/logs/SalsrvLog> file reporting how many orphans were found and
260 the approximate number of kilobytes they are consuming. This is the
261 default if the B<-orphans> argument is omitted.
265 Removes the orphaned objects, and prints a message to the
266 F</usr/afs/logs/SalsrvLog> file reporting how many orphans were removed
267 and the approximate number of kilobytes they were consuming.
271 Attaches the orphaned objects by creating a reference to them in the vnode
272 of the volume's root directory. Since each object's actual name is now
273 lost, the Salvageserver assigns each one a name of the following form:
277 =item C<__ORPHANFILE__.I<index>> for files.
279 =item C<__ORPHANDIR__.I<index>> for directories.
283 where I<index> is a two-digit number that uniquely identifies each
284 object. The orphans are charged against the volume's quota and appear in
285 the output of the B<ls> command issued against the volume's root
292 Specifies that logging output should go to syslog instead of the log file.
294 =item B<-syslogfacility> <I<Syslog facility number to use>>
296 Specify to which facility log messages should be sent when B<-syslog> is
301 Salvageserver runs in client Mode. The requested volume on the requested
302 partition will be scheduled for salvaging by the Salvageserver daemon.
304 =item B<-logfile> <I<log file>>
306 Sets the file to use for server logging. If logfile is not specified and
307 no other logging options are supplied, this will be F</usr/afs/logs/SalsrvLog>.
308 Note that this option is intended for debugging and testing purposes.
309 Changing the location of the log file from the command line may result
310 in undesirable interactions with tools such as B<bos>.
314 Prints the online help for this command. All other valid options are
321 The following command instructs the Salvageserver to schedule the salvage
322 of the volume with volume ID 258347486 on F</vicepg> on the local machine.
324 % /usr/afs/bin/salvageserver -partition /vicepg -volumeid 258347486 -client
326 =head1 PRIVILEGE REQUIRED
328 To issue the command at the shell prompt, the issuer must be logged in as
329 the local superuser C<root>.
343 IBM Corporation 2000. <http://www.ibm.com/> All Rights Reserved.
344 Sine Nomine Associates 2008. All Rights Reserved.
346 This documentation is covered by the IBM Public License Version 1.0. It was
347 converted from HTML to POD by software written by Chas Williams and Russ
348 Allbery, based on work by Alf Wachsmann and Elizabeth Cassell. This document
349 was adapted from the Salvager POD documentation.