3 fileserver - Initializes the File Server component of the fs process
10 B<fileserver> S<<< [B<-auditlog> <I<path to log file>>] >>>
11 S<<< [B<-d> <I<debug level>>] >>>
12 S<<< [B<-p> <I<number of processes>>] >>>
13 S<<< [B<-spare> <I<number of spare blocks>>] >>>
14 S<<< [B<-pctspare> <I<percentage spare>>] >>> S<<< [B<-b> <I<buffers>>] >>>
15 S<<< [B<-l> <I<large vnodes>>] >>> S<<< [B<-s> <I<small vnodes>>] >>>
16 S<<< [B<-vc> <I<volume cachesize>>] >>> S<<< [B<-w> <I<call back wait interval>>] >>>
17 S<<< [B<-cb> <I<number of call backs>>] >>> [B<-banner>] [B<-novbc>]
18 S<<< [B<-implicit> <I<admin mode bits: rlidwka>>] >>> [B<-readonly>]
19 S<<< [B<-hr> <I<number of hours between refreshing the host cps>>] >>>
20 S<<< [B<-busyat> <I<< redirect clients when queue > n >>>] >>>
21 [B<-nobusy>] S<<< [B<-rxpck> <I<number of rx extra packets>>] >>>
22 [B<-rxdbg>] [B<-rxdbge>] S<<< [B<-rxmaxmtu> <I<bytes>>] >>>
23 [B<-allow-dotted-principal>]
24 S<<< [B<-rxbind> <I<address to bind the Rx socket to>>] >>>
25 S<<< [B<-vattachpar> <I<number of volume attach threads>>] >>>
26 S<<< [B<-m> <I<min percentage spare in partition>>] >>>
27 [B<-lock>] [B<-L>] [B<-S>] S<<< [B<-k> <I<stack size>>] >>>
28 S<<< [B<-realm> <I<Kerberos realm name>>] >>>
29 S<<< [B<-udpsize> <I<size of socket buffer in bytes>>] >>>
30 S<<< [B<-sendsize> <I<size of send buffer in bytes>>] >>>
31 S<<< [B<-abortthreshold> <I<abort threshold>>] >>>
32 [B<-enable_peer_stats>] [B<-enable_process_stats>] [B<-help>]
39 The B<fileserver> command initializes the File Server component of the
40 C<fs> process. In the conventional configuration, its binary file is
41 located in the F</usr/afs/bin> directory on a file server machine.
43 The B<fileserver> command is not normally issued at the command shell
44 prompt, but rather placed into a database server machine's
45 F</usr/afs/local/BosConfig> file with the B<bos create> command. If it is
46 ever issued at the command shell prompt, the issuer must be logged onto a
47 file server machine as the local superuser C<root>.
49 The File Server creates the F</usr/afs/logs/FileLog> log file as it
50 initializes, if the file does not already exist. It does not write a
51 detailed trace by default, but the B<-d> option may be used to
52 increase the amount of detail. Use the B<bos getlog> command to
53 display the contents of the log file.
55 The command's arguments enable the administrator to control many aspects
56 of the File Server's performance, as detailed in L<OPTIONS>. By default
57 the B<fileserver> command sets values for many arguments that are suitable
58 for a medium-sized file server machine. To set values suitable for a small
59 or large file server machine, use the B<-S> or B<-L> flag
60 respectively. The following list describes the parameters and
61 corresponding argument for which the B<fileserver> command sets default
62 values, and the table below summarizes the setting for each of the three
69 The maximum number of lightweight processes (LWPs) the File Server uses to
70 handle requests for data; corresponds to the B<-p> argument. The File
71 Server always uses a minimum of 32 KB of memory for these processes.
75 The maximum number of directory blocks the File Server caches in memory;
76 corresponds to the B<-b> argument. Each cached directory block (buffer)
77 consumes 2,092 bytes of memory.
81 The maximum number of large vnodes the File Server caches in memory for
82 tracking directory elements; corresponds to the B<-l> argument. Each large
83 vnode consumes 292 bytes of memory.
87 The maximum number of small vnodes the File Server caches in memory for
88 tracking file elements; corresponds to the B<-s> argument. Each small
89 vnode consumes 100 bytes of memory.
93 The maximum volume cache size, which determines how many volumes the File
94 Server can cache in memory before having to retrieve data from disk;
95 corresponds to the B<-vc> argument.
99 The maximum number of callback structures the File Server caches in
100 memory; corresponds to the B<-cb> argument. Each callback structure
101 consumes 16 bytes of memory.
105 The maximum number of Rx packets the File Server uses; corresponds to the
106 B<-rxpck> argument. Each packet consumes 1544 bytes of memory.
110 The default values are:
112 Parameter (Argument) Small (-S) Medium Large (-L)
113 ---------------------------------------------------------------------
114 Number of LWPs (-p) 6 9 12
115 Number of cached dir blocks (-b) 70 90 120
116 Number of cached large vnodes (-l) 200 400 600
117 Number of cached small vnodes (-s) 200 400 600
118 Maximum volume cache size (-vc) 200 400 600
119 Number of callbacks (-cb) 20,000 60,000 64,000
120 Number of Rx packets (-rxpck) 100 150 200
122 To override any of the values, provide the indicated argument (which can
123 be combined with the B<-S> or B<-L> flag).
125 The amount of memory required for the File Server varies. The approximate
126 default memory usage is 751 KB when the B<-S> flag is used (small
127 configuration), 1.1 MB when all defaults are used (medium configuration),
128 and 1.4 MB when the B<-L> flag is used (large configuration). If
129 additional memory is available, increasing the value of the B<-cb> and
130 B<-vc> arguments can improve File Server performance most directly.
132 By default, the File Server allows a volume to exceed its quota by 1 MB
133 when an application is writing data to an existing file in a volume that
134 is full. The File Server still does not allow users to create new files in
135 a full volume. To change the default, use one of the following arguments:
141 Set the B<-spare> argument to the number of extra kilobytes that the File
142 Server allows as overage. A value of C<0> allows no overage.
146 Set the B<-pctspare> argument to the percentage of the volume's quota the
147 File Server allows as overage.
151 By default, the File Server implicitly grants the C<a> (administer) and
152 C<l> (lookup) permissions to system:administrators on the access control
153 list (ACL) of every directory in the volumes stored on its file server
154 machine. In other words, the group's members can exercise those two
155 permissions even when an entry for the group does not appear on an ACL. To
156 change the set of default permissions, use the B<-implicit> argument.
158 The File Server maintains a I<host current protection subgroup> (I<host
159 CPS>) for each client machine from which it has received a data access
160 request. Like the CPS for a user, a host CPS lists all of the Protection
161 Database groups to which the machine belongs, and the File Server compares
162 the host CPS to a directory's ACL to determine in what manner users on the
163 machine are authorized to access the directory's contents. When the B<pts
164 adduser> or B<pts removeuser> command is used to change the groups to
165 which a machine belongs, the File Server must recompute the machine's host
166 CPS in order to notice the change. By default, the File Server contacts
167 the Protection Server every two hours to recompute host CPSs, implying
168 that it can take that long for changed group memberships to become
169 effective. To change this frequency, use the B<-hr> argument.
171 The File Server stores volumes in partitions. A partition is a
172 filesystem or directory on the server machine that is named C</vicepX>
173 or C</vicepXX> where XX is "a" through "z" or "aa" though "zz". The
174 File Server expects that the /vicepXX directories are each on a
175 dedicated filesystem. The File Server will only use a /vicepXX if it's
176 a mountpoint for another filesystem, unless the file
177 C</vicepXX/AlwaysAttach> exists. The data in the partition is a
178 special format that can only be access using OpenAFS commands or an
181 The File Server generates the following message when a partition is nearly
184 No space left on device
186 This command does not use the syntax conventions of the AFS command
187 suites. Provide the command name and all option names in full.
191 Do not use the B<-k> and B<-w> arguments, which are intended for use
192 by the OpenAFS developers only. Changing them from their default
193 values can result in unpredictable File Server behavior. In any case,
194 on many operating systems the File Server uses native threads rather
195 than the LWP threads, so using the B<-k> argument to set the number of
196 LWP threads has no effect.
198 Do not specify both the B<-spare> and B<-pctspare> arguments. Doing so
199 causes the File Server to exit, leaving an error message in the
200 F</usr/afs/logs/FileLog> file.
202 Options that are available only on some system types, such as the B<-m>
203 and B<-lock> options, appear in the output generated by the B<-help>
204 option only on the relevant system type.
206 Currently, the maximum size of a volume is 2 terabytes (2^31 bytes)
207 and the maximum size of a /vicepX partition on a fileserver is also 2
208 terabytes. The fileserver will not report an error when it has access
209 to a partition larger than 2 terabytes, but it will probably fail if
210 the administrator attempts to use more than 2 terabytes of space. In
211 addition, there are reports of erroneous disk usage numbers when
212 B<vos partinfo> or other OpenAFS disk reporting tools are used with
213 partitions larger than 2 terabytes.
215 The maximum number of directory entries is 64,000 if all of the
216 entries have names that are 15 characters or less in length. A name
217 that is 15 characters long requires the use of only one block in the
218 directory. Additional sequential blocks are required to store entries
219 with names that are longer than 15 characters. Each additional block
220 provides an additional length of 32 characters for the name of the
223 In real world use, the maximum number of objects in an AFS directory
224 is usually between 16,000 and 25,000, depending on the average name
231 =item B<-auditlog> <I<log path>>
233 Set and enable auditing.
235 =item B<-d> <I<debug level>>
237 Sets the detail level for the debugging trace written to the
238 F</usr/afs/logs/FileLog> file. Provide one of the following values, each
239 of which produces an increasingly detailed trace: C<0>, C<1>, C<5>, C<25>,
240 and C<125>. The default value of C<0> produces only a few messages.
242 =item B<-p> <I<number of processes>>
244 Sets the number of threads to run. Provide a positive integer. The File
245 Server creates and uses five threads for special purposes, in addition to
246 the number specified (but if this argument specifies the maximum possible
247 number, the File Server automatically uses five of the threads for its own
250 The maximum number of threads can differ in each release of AFS. Consult
251 the I<IBM AFS Release Notes> for the current release.
253 =item B<-spare> <I<number of spare blocks>>
255 Specifies the number of additional kilobytes an application can store in a
256 volume after the quota is exceeded. Provide a positive integer; a value of
257 C<0> prevents the volume from ever exceeding its quota. Do not combine
258 this argument with the B<-pctspare> argument.
260 =item B<-pctspare> <I<percentage spare>>
262 Specifies the amount by which the File Server allows a volume to exceed
263 its quota, as a percentage of the quota. Provide an integer between C<0>
264 and C<99>. A value of C<0> prevents the volume from ever exceeding its
265 quota. Do not combine this argument with the B<-spare> argument.
267 =item B<-b> <I<buffers>>
269 Sets the number of directory buffers. Provide a positive integer.
271 =item B<-l> <I<large vnodes>>
273 Sets the number of large vnodes available in memory for caching directory
274 elements. Provide a positive integer.
276 =item B<-s> <I<small nodes>>
278 Sets the number of small vnodes available in memory for caching file
279 elements. Provide a positive integer.
281 =item B<-vc> <I<volume cachesize>>
283 Sets the number of volumes the File Server can cache in memory. Provide a
286 =item B<-w> <I<call back wait interval>>
288 Sets the interval at which the daemon spawned by the File Server performs
289 its maintenance tasks. Do not use this argument; changing the default
290 value can cause unpredictable behavior.
292 =item B<-cb> <I<number of callbacks>>
294 Sets the number of callbacks the File Server can track. Provide a positive
299 Prints the following banner to F</dev/console> about every 10 minutes.
301 File Server is running at I<time>.
305 Prevents the File Server from breaking the callbacks that Cache Managers
306 hold on a volume that the File Server is reattaching after the volume was
307 offline (as a result of the B<vos restore> command, for example). Use of
308 this flag is strongly discouraged.
310 =item B<-implicit> <I<admin mode bits>>
312 Defines the set of permissions granted by default to the
313 system:administrators group on the ACL of every directory in a volume
314 stored on the file server machine. Provide one or more of the standard
315 permission letters (C<rlidwka>) and auxiliary permission letters
316 (C<ABCDEFGH>), or one of the shorthand notations for groups of permissions
317 (C<all>, C<none>, C<read>, and C<write>). To review the meaning of the
318 permissions, see the B<fs setacl> reference page.
320 =item B<-hr> <I<number of hours between refreshing the host cps>>
322 Specifies how often the File Server refreshes its knowledge of the
323 machines that belong to protection groups (refreshes the host CPSs for
324 machines). The File Server must update this information to enable users
325 from machines recently added to protection groups to access data for which
326 those machines now have the necessary ACL permissions.
328 =item B<-busyat> <I<< redirect clients when queue > n >>>
330 Defines the number of incoming RPCs that can be waiting for a response
331 from the File Server before the File Server returns the error code
332 C<VBUSY> to the Cache Manager that sent the latest RPC. In response, the
333 Cache Manager retransmits the RPC after a delay. This argument prevents
334 the accumulation of so many waiting RPCs that the File Server can never
335 process them all. Provide a positive integer. The default value is
338 =item B<-rxpck> <I<number of rx extra packets>>
340 Controls the number of Rx packets the File Server uses to store data for
341 incoming RPCs that it is currently handling, that are waiting for a
342 response, and for replies that are not yet complete. Provide a positive
347 Writes a trace of the File Server's operations on Rx packets to the file
348 F</usr/afs/logs/rx_dbg>.
352 Writes a trace of the File Server's operations on Rx events (such as
353 retransmissions) to the file F</usr/afs/logs/rx_dbg>.
355 =item B<-allow-dotted-principal>
357 By default, the RXKAD security layer will disallow access by Kerberos
358 principals with a dot in the first component of their name. This is to avoid
359 the confusion where principals user/admin and user.admin are both mapped to the
360 user.admin PTS entry. Sites whose Kerberos realms don't have these collisions
361 between principal names may disable this check by starting the server
364 =item F<-m> <I<min percentage spare in partition>>
366 Specifies the percentage of each AFS server partition that the AIX version
367 of the File Server creates as a reserve. Specify an integer value between
368 C<0> and C<30>; the default is 8%. A value of C<0> means that the
369 partition can become completely full, which can have serious negative
374 Prevents any portion of the fileserver binary from being paged (swapped)
375 out of memory on a file server machine running the IRIX operating system.
379 Sets values for many arguments in a manner suitable for a large file
380 server machine. Combine this flag with any option except the B<-S> flag;
381 omit both flags to set values suitable for a medium-sized file server
386 Sets values for many arguments in a manner suitable for a small file
387 server machine. Combine this flag with any option except the B<-L> flag;
388 omit both flags to set values suitable for a medium-sized file server
391 =item B<-k> <I<stack size>>
393 Sets the LWP stack size in units of 1 kilobyte. Do not use this argument,
394 and in particular do not specify a value less than the default of C<24>.
396 =item B<-realm> <I<Kerberos realm name>>
398 Defines the Kerberos realm name for the File Server to use. If this
399 argument is not provided, it uses the realm name corresponding to the cell
400 listed in the local F</usr/afs/etc/ThisCell> file.
402 =item B<-udpsize> <I<size of socket buffer in bytes>>
404 Sets the size of the UDP buffer, which is 64 KB by default. Provide a
405 positive integer, preferably larger than the default.
407 =item B<-enable_peer_stats>
409 Activates the collection of Rx statistics and allocates memory for their
410 storage. For each connection with a specific UDP port on another machine,
411 a separate record is kept for each type of RPC (FetchFile, GetStatus, and
412 so on) sent or received. To display or otherwise access the records, use
413 the Rx Monitoring API.
415 =item B<-enable_process_stats>
417 Activates the collection of Rx statistics and allocates memory for their
418 storage. A separate record is kept for each type of RPC (FetchFile,
419 GetStatus, and so on) sent or received, aggregated over all connections to
420 other machines. To display or otherwise access the records, use the Rx
423 =item B<-abortthreshold> <I<abort threshold>>
425 Sets the abort threshold, which is triggered when an AFS client sends
426 a number of FetchStatus requests in a row and all of them fail due to
427 access control or some other error. When the abort threshold is
428 reached, the file server starts to slow down the responses to the
429 problem client in order to reduce the load on the file server.
431 The throttling behaviour can cause issues especially for some versions
432 of the Windows OpenAFS client. When using Windows Explorer to navigate
433 the AFS directory tree, directories with only "look" access for the
434 current user may load more slowly because of the throttling. This is
435 because the Windows OpenAFS client sends FetchStatus calls one at a
436 time instead of in bulk like the Unix Open AFS client.
438 Setting the threshold to 0 disables the throttling behavior. This
439 option is available in OpenAFS versions 1.4.1 and later.
443 Prints the online help for this command. All other valid options are
450 The following B<bos create> command creates an fs process on the file
451 server machine C<fs2.abc.com> that uses the large configuration size, and
452 allows volumes to exceed their quota by 10%. Type the command on a single
455 % bos create -server fs2.abc.com -instance fs -type fs \
456 -cmd "/usr/afs/bin/fileserver -pctspare 10 \
457 -L" /usr/afs/bin/volserver /usr/afs/bin/salvager
460 =head1 TROUBLESHOOTING
462 Sending process signals to the File Server Process can change its
463 behavior in the following ways:
465 Process Signal OS Result
466 ---------------------------------------------------------------------
468 File Server XCPU Unix Prints a list of client IP
471 File Server USR2 Windows Prints a list of client IP
474 File Server POLL HPUX Prints a list of client IP
477 Any server TSTP Any Increases Debug level by a power
478 of 5 -- 1,5,25,125, etc.
479 This has the same effect as the
480 -d XXX command-line option.
482 Any Server HUP Any Resets Debug level to 0
484 File Server TERM Any Run minor instrumentation over
485 the list of descriptors.
487 Other Servers TERM Any Causes the process to quit.
489 File Server QUIT Any Causes the File Server to Quit.
490 Bos Server knows this.
492 The basic metric of whether an AFS file server is doing well is the number
493 of connections waiting for a thread,
494 which can be found by running the following command:
496 % rxdebug <server> | grep waiting_for | wc -l
498 Each line returned by C<rxdebug> that contains the text "waiting_for"
499 represents a connection that's waiting for a file server thread.
501 If the blocked connection count is ever above 0, the server is having
502 problems replying to clients in a timely fashion. If it gets above 10,
503 roughly, there will be noticable slowness by the user. The total number of
504 connections is a mostly irrelevant number that goes essentially
505 monotonically for as long as the server has been running and then goes back
506 down to zero when it's restarted.
508 The most common cause of blocked connections rising on a server is some
509 process somewhere performing an abnormal number of accesses to that server
510 and its volumes. If multiple servers have a blocked connection count, the
511 most likely explanation is that there is a volume replicated between those
512 servers that is absorbing an abnormally high access rate.
514 To get an access count on all the volumes on a server, run:
516 % vos listvol <server> -long
518 and save the output in a file. The results will look like a bunch of B<vos
519 examine> output for each volume on the server. Look for lines like:
521 40065 accesses in the past day (i.e., vnode references)
523 and look for volumes with an abnormally high number of accesses. Anything
524 over 10,000 is fairly high, but some volumes like root.cell and other
525 volumes close to the root of the cell will have that many hits routinely.
526 Anything over 100,000 is generally abnormally high. The count resets about
529 Another approach that can be used to narrow the possibilities for a
530 replicated volume, when multiple servers are having trouble, is to find all
531 replicated volumes for that server. Run:
533 % vos listvldb -server <server>
535 where <server> is one of the servers having problems to refresh the VLDB
538 % vos listvldb -server <server> -part <partition>
540 to get a list of all volumes on that server and partition, including every
541 other server with replicas.
543 Once the volume causing the problem has been identified, the best way to
544 deal with the problem is to move that volume to another server with a low
545 load or to stop any runaway programs that are accessing that volume
546 unnecessarily. Often the volume will be enough information to tell what's
549 If you still need additional information about who's hitting that server,
550 sometimes you can guess at that information from the failed callbacks in the
551 F<FileLog> log in F</var/log/afs> on the server, or from the output of:
553 % /usr/afsws/etc/rxdebug <server> -rxstats
555 but the best way is to turn on debugging output from the file server.
556 (Warning: This generates a lot of output into FileLog on the AFS server.)
557 To do this, log on to the AFS server, find the PID of the fileserver
562 where <pid> is the PID of the file server process. This will raise the
563 debugging level so that you'll start seeing what people are actually doing
564 on the server. You can do this up to three more times to get even more
565 output if needed. To reset the debugging level back to normal, use (The
566 following command will NOT terminate the file server):
570 The debugging setting on the File Server should be reset back to normal when
571 debugging is no longer needed. Otherwise, the AFS server may well fill its
572 disks with debugging output.
574 The lines of the debugging output that are most useful for debugging load
577 SAFS_FetchStatus, Fid = 2003828163.77154.82248, Host 171.64.15.76
578 SRXAFS_FetchData, Fid = 2003828163.77154.82248
580 (The example above is partly truncated to highlight the interesting
581 information). The Fid identifies the volume and inode within the volume;
582 the volume is the first long number. So, for example, this was:
584 % vos examine 2003828163
585 pubsw.matlab61 2003828163 RW 1040060 K On-line
586 afssvr5.Stanford.EDU /vicepa
587 RWrite 2003828163 ROnly 2003828164 Backup 2003828165
589 Creation Mon Aug 6 16:40:55 2001
590 Last Update Tue Jul 30 19:00:25 2002
591 86181 accesses in the past day (i.e., vnode references)
593 RWrite: 2003828163 ROnly: 2003828164 Backup: 2003828165
595 server afssvr5.Stanford.EDU partition /vicepa RW Site
596 server afssvr11.Stanford.EDU partition /vicepd RO Site
597 server afssvr5.Stanford.EDU partition /vicepa RO Site
599 and from the Host information one can tell what system is accessing that
602 Note that the output of L<vos_examine(1)> also includes the access count, so
603 once the problem has been identified, vos examine can be used to see if the
604 access count is still increasing. Also remember that you can run vos
605 examine on the read-only replica (e.g., pubsw.matlab61.readonly) to see the
606 access counts on the read-only replica on all of the servers that it's
609 =head1 PRIVILEGE REQUIRED
611 The issuer must be logged in as the superuser C<root> on a file server
612 machine to issue the command at a command shell prompt. It is conventional
613 instead to create and start the process by issuing the B<bos create>
629 IBM Corporation 2000. <http://www.ibm.com/> All Rights Reserved.
631 This documentation is covered by the IBM Public License Version 1.0. It was
632 converted from HTML to POD by software written by Chas Williams and Russ
633 Allbery, based on work by Alf Wachsmann and Elizabeth Cassell.