1 Amanda 2.4 - System-Specific Installation Notes
3 Please read the notes that correspond to the architectures you are
4 installing for. If you find additional gotchas, or anything incorrect
5 in these notes, please send your updates to amanda-hackers@amanda.org,
6 after checking that they are not known/fixed problems in the amanda
7 patches page: http://www.amanda.org/patches.html
9 ---------------------------------------------------------------------------
13 You may have compilation errors on statfs.c if you're running, on a
14 Solaris 2.6 host, a gcc that was not build on a Solaris 2.6 host.
15 This happens because gcc stores fixed copies of some Solaris header
16 files on an internal directory. You must rebuild gcc if you get this
17 kind of trouble. Note, however, that gcc 2.7.2.3 does not support
18 Solaris 2.6, you should upgrade to 2.8.0 or higher, or egcs.
20 ---------------------------------------------------------------------------
24 According to Julian Stevens <julian.stevens@baedsl.co.uk>, the format
25 of inetd.conf on Trusted Solaris 1.2 is a bit different. Just before
26 the user name, you should insert the word `untrusted':
28 amanda dgram udp wait untrusted amuser /usr/local/libexec/amandad amandad
30 The patch-system script is *NOT* aware of this detail; you must fix it
33 ---------------------------------------------------------------------------
37 A bug in GNU tar 1.12 causes it to miscalculate (in fact, to
38 misreport) the size of filesystems. A patch for GNU tar is available
39 in the patches directory.
41 ---------------------------------------------------------------------------
45 The Ultrix dump program contains an explicit check that it is being
46 run by root. This defeats the usual practice of a non-root "operator"
47 userid for dumps. For this reason, the rundump program (a setuid-root
48 wrapper for dump) is enabled by default. If you find rundump is not
49 necessary for you, just configure --without-rundump.
52 The Ultrix restore program will fail if it is not connected to a tty.
53 Since the restore program is invoked in the clients in order to create
54 index files, and it the client is not connected to a tty, index
55 creation will fail. Using GNUTAR instead of DUMP is an option.
56 Thanks to Edwin Wiles <ewiles@mclean.sterling.com> for the
57 investigation. Another alternative proposed by Martyn Johnson
58 <Martyn.Johnson@cl.cam.ac.uk> is to use a modified restore program:
59 use a binary program editor and replace `/dev/tty' with `/dev/nul',
60 for instance, and link /dev/nul to /dev/null. Note that the chosen
61 file name must be exactly 8 bytes long, otherwise you'll break the
62 restore program. A nice one-liner perl script by Martyn Johnson will
63 do the trick (make sure you preserve a copy of the original restore
64 program, it will be rewritten by running this script!):
66 perl -pi -e 'BEGIN { $/ = "/dev/tty" } s-$/-/dev/nul-' restore
69 The Ultrix C compiler seems to be too broken to compile Amanda. You
70 are likely to need gcc or egcs.
72 ---------------------------------------------------------------------------
76 You may run into an internal /bin/sh limit when running the configure
77 script. The error message is:
79 ./configure: sh internal 2K buffer overflow
81 Using /bin/posix/sh usually works around the problem. Another
82 solution is to install GNU bash and use it instead of /bin/sh.
85 If `configure' complains about not finding `lex', you'll have to get
86 `flex' installed. Look for its URL in docs/INSTALL.
89 If you use logical volumes, you may refer to mountpoints or full
90 device pathnames instead of device names in the disk list file.
93 According to Stan Brown <stanb@awod.com>, amverify won't work with
94 HP/UX's stock `mt'. The work-around is to install GNU cpio, that
95 contains an implementation of `mt', and edit amverify so that MT
96 points to GNU mt and MTOPT is `-f', or reconfigure and rebuild amanda
97 from scratch, making sure it finds GNU mt before the stock mt in the
101 If you have vxfs filesystems to back up, amanda will pick vxdump
105 GNU tar 1.12 will incorrectly report the size of backups. There's a
106 patch in the patches directory that fixes this problem.
109 The use of `amhpfixdevs' is deprecated, since you can list
110 mount-points or full device names in the disklist. The script may be
111 removed in future releases of Amanda.
114 Sometimes you may get the error message `Link severed' from an HP/UX
115 client. This is just a cute name the developers of HP/UX found for
116 `Network error'. Reported by Michael Brehl <mbr@condor.de>
118 ---------------------------------------------------------------------------
122 Linux hosts intended to back up efs partitions with dump should
123 install the dump package, as it is not installed by default on most
124 Linux distributions. It is possible to find compiled versions of dump
125 on most Linux sites and CD-ROMs.
127 Note, however, that DUMP for Linux has been quite unreliable,
128 sometimes being unable to perform backups, producing weird error
129 messages such as `master/slave protocol botched', and sometimes
130 creating unrestorable dump images, especially of active filesystems.
131 The first problem can sometimes be fixed by cleaning up outdated
132 entries in /etc/dumpdates, but your best bet is probably GNU tar.
135 Make sure the user that runs configure has permission to run the dump
136 program, otherwise configure may misdetect an ability of dump to
137 accept a -E (for estimates) switch.
140 GNUtar 1.11.8, distributed with some Linux versions, will cause index
141 failures (Index returned -1). Upgrading to GNUtar 1.12 fixes this
142 problem. This is not a Linux-specific problem, but it is quite common
146 Amanda now supports the ftape driver version 3.04d. It adjusts the
147 blocksize automatically to 32k and supports QIC volume tables. More
148 details con be found in the file ZFTAPE in this directory.
151 Some releases of dump for Linux, such as the one distributed with
152 Debian 2.0, have modified dump so that it stores dumpdates in
153 /var/lib. If this is your case, you should create a link to it in
154 /etc. Suggested by David Wolfskill <dhw@whistle.com>
156 ---------------------------------------------------------------------------
160 According to Michael Galloway <mgx@spruce.lsd.ornl.gov>, the stock
161 DUX4 dump is broken. There is a patch available at
162 ftp://ftp.service.digital.com/public/dunix/v4.0b/duv40bas00005-19970926.README
165 When both dump and vdump are available, amanda will use vdump for
166 backing up advfs filesystems only, and dump will be used for the rest.
167 If you'd rather back up all filesystems with vdump, #undef DUMP in
168 config/config.h after running configure.
171 Unfortunately, the output of `dump -E' incorrectly matches a line of
172 output from SAMBA, which gets Amanda's estimate process confused.
173 client-src/sendsize.c will refuse to compile if both
174 HAVE_DUMP_ESTIMATE and SAMBA_CLIENT are defined in config/config.h.
175 Amanda will work correctly if you undefine HAVE_DUMP_ESTIMATE in
176 config/config.h; if you prefer to have incorrect estimates for SAMBA
177 backups, follow the instructions in sendsize.c on removing the
181 According to Oren Laadan <orenl@cs.huji.ac.il>, DEC compiler version
182 "DEC C V5.2-033 on Digital UNIX V4.0 (Rev. 564)" (obtained with "cc -V")
183 does not build Amanda properly, in particular, taper.c. Using gcc is OK.
185 ---------------------------------------------------------------------------
186 Sinix 5.43 (Reliant Unix)
187 -------------------------
189 The use of `amsinixfixdevs' is deprecated, since you can list
190 mount-points or full device names in the disklist. The script may be
191 removed in future releases of Amanda.
193 Sinix port originally by Michael Schmitz <mschmitz@iname.com>.
195 ---------------------------------------------------------------------------
199 When setting the tape device name in either amanda.conf or one of the
200 changer configuration files, make sure you specify the "variable" device
201 name, which has a 'v' on the end. If not, IRIX will write 4KByte blocks
202 instead of the 32KByte blocks Amanda tells it to. This apparantly works
203 OK unless you take the tape to a non-IRIX system, where amrestore will
204 complain about a short (4096) read.
206 If you do end up in this situation, the dd command to use to pipe into
207 your system restore program is:
209 dd if=/dev/whatever bs=4k skip=8 | ...
212 Jean-Francois Malouin <Jean-Francois.Malouin@bic.mni.mcgill.ca>
213 reports that, if you are going to use an IRIX host as the tape server,
214 you *must* patch your system with *all* kernel and scsi rollup
215 patches, otherwise you may end up with useless tapes, due to tape
216 rewinding after short periods of inactivity. See
217 http://www-viz.tamu.edu/~sgi-faq/ for more details.
220 Some unpatched version of xfsdump are reported as not printing
221 estimates. This causes estimates to take much longer than they had
222 to, because backups are always performed twice. According to Mike
223 Acar <mike@kentinfoworks.com>, patch 2333 for IRIX 5.3 and 6.1 will
226 ---------------------------------------------------------------------------
230 Seems like SGI `make' program is a bit broken, in a way that causes it
231 to rebuild some files if doesn't have to if you happen to run `make'
232 again. Using GNU make fixes this problem.
235 If you have xfs filesystems to back up, amanda will pick xfsdump
239 ---------------------------------------------------------------------------
243 Jens Krause <jens@transcom.de> has reported some problems with GNU tar
244 1.12 on SCO Release 5. Although the `sparse' files patch was applied,
245 GNU tar would consistently crash. GNU tar had to be built linked with
246 malloc-libraries, and the `--sparse' switches had to be removed from
247 client-src/sendbackup-gnutar.c and client-src/sendsize.c.
250 ---------------------------------------------------------------------------
254 chg-scsi was not updated to support the new camlib.h-dependent chio.h,
255 so chg-scsi will be automatically disabled if camlib.h is found. You
256 may use chg-chio instead.
259 ---------------------------------------------------------------------------
263 Nicolas Ecarnot <nicolas.ecarnot@accim.com> discovered that for
264 FreeBSD 5.1 (maybe earlier, and surely further), you have to set the
265 net.inet.udp.maxdgram TCP/IP variable to 65535. The default is 9216,
266 and this is a problem when trying to backup a large number of clients
267 as indicated by errors in during amcheck or the estimate phase.
269 You can just run the command:
271 sysctl net.inet.udp.maxdgram=63535
273 but this won't last until the next reboot.
275 To make it permanent, just add this line:
277 net.inet.udp.maxdgram=65535
279 in the /etc/sysctl.conf file.
282 ---------------------------------------------------------------------------
286 sendsize is reported to coredump on AIX 4.3.3, this is a linking problem,
287 try configuring amanda with --disable-shared option.
290 ---------------------------------------------------------------------------
294 Although amanda won't run standalone on MS-Windows hosts, it is
295 possible to use it to back up their disks, by using SAMBA. Please
296 read docs/SAMBA for more information.
298 SAMBA may be unable to back up some files due to file locking
299 restrictions. Particularly, paging and registry files usually present
300 problems. Backing up page files is pointless, but registry files are
301 quite important to back up. It is possible to create regular files
302 that contain registry information by using the Regback utility, from
303 the Windows NT Resource Kit. Unfortunately, is not part of the
304 Windows NT standard distribution, you have to purchase it separately.
305 Thanks to Ernie Oporto <ernie_oporto@mentorg.com> for the tip.
projects / debian / amanda / blob