How to read a damaged tape
Hi, OK, this is a bit out of topic, but I really don't know where to ask. I have a SLR 100 tape drive (SCSI interface) and a tape cartridge with 4 files (not an Amanad tape, just plain tar). The first file is damaged so I cannot mt fsf over that file to read the others. Is there a secret way to to advance the tape by a certain number of meters, or by a certain number of seconds or something. I know I would end up at a random position in the middle of a file, but at least I could fast forward over the damaged file. Best regards, Olivier
problem backing up Windows host
I've just inherited an amanda setup that has apparently never worked properly. The specific problem I'm trying to sort at the moment is summed up below. Amanda Backup Client Hosts Check ERROR: NAK finn.internal.domain.com: access as amanda not allowed from [EMAIL PROTECTED]: cannot open /usr/amanda/etc/amanda/.amandahosts: Permission denied Client check: 1 host checked in 0.002 seconds, 1 problem found finn is a Windows server, and as far as I can see the client is setup the same on this as on another windows machine that is backing up ok. Its got 'amanda' and 'amanda_backup' as users of the machines, amanda is installed in c:\usr on both. There's a .amandahosts file on both with the same in: orinoco x.x.x.x orinoco.internal.domain.com x.x.x.x The only difference I can see is that finn is win 2000, and the one that works is win 2003. Any ideas anyone? Thanks, John
Re: problem backing up Windows host
On 2006-04-21 13:28, John Clement wrote: I've just inherited an amanda setup that has apparently never worked properly. The specific problem I'm trying to sort at the moment is summed up below. Amanda Backup Client Hosts Check ERROR: NAK finn.internal.domain.com: access as amanda not allowed from [EMAIL PROTECTED]: cannot open /usr/amanda/etc/amanda/.amandahosts: Permission denied The user amanda should be able to read that file. (and the all the upper directories too). Client check: 1 host checked in 0.002 seconds, 1 problem found finn is a Windows server, and as far as I can see the client is setup the same on this as on another windows machine that is backing up ok. Its got 'amanda' and 'amanda_backup' as users of the machines, amanda is installed in c:\usr on both. There's a .amandahosts file on both with the same in: orinoco x.x.x.x orinoco.internal.domain.com x.x.x.x This seems to imply that x.x.x.x is an ip-number. No it is not. The second item on the line is the remoteuser (defaults to local user if empty). The only difference I can see is that finn is win 2000, and the one that works is win 2003. Any ideas anyone? -- Paul Bijnens, xplanation Technology ServicesTel +32 16 397.511 Technologielaan 21 bus 2, B-3001 Leuven, BELGIUMFax +32 16 397.512 http://www.xplanation.com/ email: [EMAIL PROTECTED] *** * I think I've got the hang of it now: exit, ^D, ^C, ^\, ^Z, ^Q, ^^, * * F6, quit, ZZ, :q, :q!, M-Z, ^X^C, logoff, logout, close, bye, /bye, * * stop, end, F3, ~., ^]c, +++ ATH, disconnect, halt, abort, hangup, * * PF4, F20, ^X^X, :D::D, KJOB, F14-f-e, F8-e, kill -1 $$, shutdown, * * init 0, kill -9 1, Alt-F4, Ctrl-Alt-Del, AltGr-NumLock, Stop-A, ... * * ... Are you sure? ... YES ... Phew ... I'm out * ***
Re: problem backing up Windows host
Paul Bijnens wrote: On 2006-04-21 13:28, John Clement wrote: I've just inherited an amanda setup that has apparently never worked properly. The specific problem I'm trying to sort at the moment is summed up below. Amanda Backup Client Hosts Check ERROR: NAK finn.internal.domain.com: access as amanda not allowed from [EMAIL PROTECTED]: cannot open /usr/amanda/etc/amanda/.amandahosts: Permission denied The user amanda should be able to read that file. (and the all the upper directories too). Just checked, 'everyone' has access to that file and its parent directories... Client check: 1 host checked in 0.002 seconds, 1 problem found finn is a Windows server, and as far as I can see the client is setup the same on this as on another windows machine that is backing up ok. Its got 'amanda' and 'amanda_backup' as users of the machines, amanda is installed in c:\usr on both. There's a .amandahosts file on both with the same in: orinoco x.x.x.x orinoco.internal.domain.com x.x.x.x This seems to imply that x.x.x.x is an ip-number. No it is not. The second item on the line is the remoteuser (defaults to local user if empty). Yes, sorry, my bad, it is infact orinoco amanda orinoco.internal.domain.com amanda not sure where I got the idea it was an IP address from! Thanks, jc
Re: How to read a damaged tape
On Fri, Apr 21, 2006 at 03:36:13PM +0700, Olivier Nicole wrote: Hi, OK, this is a bit out of topic, but I really don't know where to ask. I have a SLR 100 tape drive (SCSI interface) and a tape cartridge with 4 files (not an Amanad tape, just plain tar). The first file is damaged so I cannot mt fsf over that file to read the others. Is there a secret way to to advance the tape by a certain number of meters, or by a certain number of seconds or something. I know I would end up at a random position in the middle of a file, but at least I could fast forward over the damaged file. Just ideas, YMMV. dd to devnull dd with skip x records to devnull mt asf mt eom then mt bsf -- Jon H. LaBadie [EMAIL PROTECTED] JG Computing 4455 Province Line Road(609) 252-0159 Princeton, NJ 08540-4322 (609) 683-7220 (fax)
Re: problem backing up Windows host
On Fri, Apr 21, 2006 at 12:28:05PM +0100, John Clement wrote: I've just inherited an amanda setup that has apparently never worked properly. The specific problem I'm trying to sort at the moment is summed up below. Amanda Backup Client Hosts Check ERROR: NAK finn.internal.domain.com: access as amanda not allowed from [EMAIL PROTECTED]: cannot open /usr/amanda/etc/amanda/.amandahosts: Permission denied Client check: 1 host checked in 0.002 seconds, 1 problem found finn is a Windows server, and as far as I can see the client is setup the same on this as on another windows machine that is backing up ok. Its got 'amanda' and 'amanda_backup' as users of the machines, amanda is installed in c:\usr on both. There's a .amandahosts file on both with the same in: Windows systems with a /usr ??? Are you using cygwin on the Windows clients to allow use of amanda client software directly on the client? If so, might there be some cygwin issues? Or, it says permission denied for .amandahosts. What are the access rights for user amanda to the .amandahosts file on the client? Windows clients often backed up as indirect clients with samba. In that case the direct client needs proper access from the amanda server and to the smb client. -- Jon H. LaBadie [EMAIL PROTECTED] JG Computing 4455 Province Line Road(609) 252-0159 Princeton, NJ 08540-4322 (609) 683-7220 (fax)
Re: problem backing up Windows host
On 2006-04-21 14:22, John Clement wrote: Paul Bijnens wrote: On 2006-04-21 13:28, John Clement wrote: I've just inherited an amanda setup that has apparently never worked properly. The specific problem I'm trying to sort at the moment is summed up below. Amanda Backup Client Hosts Check ERROR: NAK finn.internal.domain.com: access as amanda not allowed from [EMAIL PROTECTED]: cannot open /usr/amanda/etc/amanda/.amandahosts: Permission denied The user amanda should be able to read that file. (and the all the upper directories too). Just checked, 'everyone' has access to that file and its parent directories... Not everyone but specifically user amanda needs to have access (yes there is a difference). (...access as amanda ) Try this as user amanda, not just looking at the permissions. -- Paul Bijnens, xplanation Technology ServicesTel +32 16 397.511 Technologielaan 21 bus 2, B-3001 Leuven, BELGIUMFax +32 16 397.512 http://www.xplanation.com/ email: [EMAIL PROTECTED] *** * I think I've got the hang of it now: exit, ^D, ^C, ^\, ^Z, ^Q, ^^, * * F6, quit, ZZ, :q, :q!, M-Z, ^X^C, logoff, logout, close, bye, /bye, * * stop, end, F3, ~., ^]c, +++ ATH, disconnect, halt, abort, hangup, * * PF4, F20, ^X^X, :D::D, KJOB, F14-f-e, F8-e, kill -1 $$, shutdown, * * init 0, kill -9 1, Alt-F4, Ctrl-Alt-Del, AltGr-NumLock, Stop-A, ... * * ... Are you sure? ... YES ... Phew ... I'm out * ***
2.5.0 problems on Solaris 8
I'm starting to upgrade my clients. The HP-UX boxes (10.20) went flawlessly. But I'm having a few issues with my first Solaris 8 machine. Now in the interest of full disclosure, these machines are supported by a 3rd party OEM, and the machine I use for a build machine was recently rebuilt by them to address a kernel memory leak issue, so odds are good what I'm seeing is a result of their meddling, but I could use some help figuring this out. I built 2.5.0 last night on this machine, using the following configure script: #!/bin/ksh export PATH=/opt/FSFbinut/bin:/usr/local/bin:/bin:/usr/bin:/usr/sbin:/opt/FSFm4/ /bin:/opt/FSF/binutils/bin:/etc:/usr/local:.:/usr/opt/SUNWmd/sbin:/usr/openwin/b in:/opt/FSFless/bin:/opt/DSIelm/bin:/opt/VAlsof/bin:/opt/LWperl/bin:/opt/FSFscre en/bin:/opt/USCsysi/bin:/usr/local/bin:/opt/WLtop/bin:/opt/VJtr/bin:/opt/FSFgdb/ bin:/opt/FSFgzip/bin:/opt/FSFlibg++/bin:/opt/LYNX271/bin:/opt/UNLncftp/bin:/opt/ netscape:/opt/FSFlibg++/bin:/opt/GCC2721/bin:/opt/TWgnuplot/bin:/usr/lib/ICAClie nt:/opt/FSFgroff/bin:/opt/hpnp/bin:/opt/FSFflex/bin:/opt/ucd-snmp/bin ./configure --with-config=DailyDump \ --prefix=/opt/amanda --with-index-server=debian --with-user=amanda \ --with-group=operator --with-tape-server=debian --with-debugging \ --with-tape-device=/dev/nst0 --sbindir=/opt/amanda/sbin \ --libexecdir=/opt/amanda/libexec --libdir=/opt/amanda/lib \ --with-configdir=/opt/amanda/etc/amanda \ --with-gnutar-listdir=/opt/amanda/var/amanda/gnutar-lists \ --mandir=/opt/amanda/man --without-rundump \ --without-server My first amcheck failed. I tried to run amandad by hand, and found that the dynamic linker was failing to find libgcc_s.so.1. I looked in .cshrc and .profile, and found that they had broken LD_LIBRARY_PATH. I fixed this, restarted inetd, and was able to run an amcheck. So far so good. But, the next backups from this machine failed, and when I look in the debug file, I find this: ld.so.1: /opt/amanda/libexec/noop: fatal: libgcc_s.so.1: open failed: No such file or directory I'm a bit confused by what sets up the search path for shared libraries under Solaris 8, when they are invoked as services like Amanda does. Is there a way to give the binaries a hard coded search path? -- U.S. Encouraged by Vietnam Vote - Officials Cite 83% Turnout Despite Vietcong Terror - New York Times 9/3/1967
Re: 2.5.0 problems on Solaris 8
stan schrieb: I'm starting to upgrade my clients. The HP-UX boxes (10.20) went flawlessly. But I'm having a few issues with [...] Sorry to be the bad guy again: Please don't hijack threads, this messes up the threads etc. Read http://en.wikipedia.org/wiki/Thread_Hijacking for details. Simply write a new message to amanda-users@amanda.org to start a new thread. Thank you, Stefan.
Re: 2.5.0 problems on Solaris 8
Stefan G. Weichinger schrieb: stan schrieb: I'm starting to upgrade my clients. The HP-UX boxes (10.20) went flawlessly. But I'm having a few issues with [...] Sorry to be the bad guy again: Please don't hijack threads, this messes up the threads etc. APOLOGIES I seem to have had mixed up my display-settings. Right now it looks OK again. Sorry for blaming you, stan.
Re: Dumps performed, but also MISSING ???
Hi, I found the problem, it will report a MISSING for all dump that doesn't start the same day you run your amdump. I will provide a patch soon. Jean-Louis jeffrey d anderson wrote: The strange behavior described here started when I upgraded my server to amanda-2.5.0. I was wondering if anyone else had seen this behavior: I run amanda 2.5.0 with a Redhat Linux 9 server attached to an ADIC Scalar 24 changer. My configuration backs up about 20 clients (mix of linux and SunOS, running various amanda versions from 2.4.3 to 2.5.0) with about 65 DLEs. I backup about 200GB per night with a dumpcycle of 7 days. This configuration has been successfully running for about a year, and seems to still be successful. So here is the point of this message. I upgraded the server from amanda 2.4.4p4 to 2.5.0 on March 30 and things appear to be running fine. Backups proceed and I've done successful restores. But In the amanda mail report I get messages like this: DUMP SUMMARY: DUMPER STATS TAPER STATS HOSTNAME DISKLORIG-kB OUT-kB COMP% client1.here.org / 1 23290 5568 23.9 client1.here.org / MISSING -- client2.here.org -/coldprobe 0 93053120 93053152-- client2.here.org -/coldprobe MISSING- It looks like these dumps have succeeded, but they are ALSO reported as missing! Not all DLEs are reported missing, and the MISSING ones are different each night (different hosts, different DLEs). Some nights there are no MISSING entries, some nights there are a couple, some nights there are five or six. This started with the upgrade to amanda-2.5.0 on the server. No changes were made to the disklist at that time. I thought perhaps I had duplicate entries in the disklist file, and amanda's handling of them had changed, but that does not appear to be the case. Is this something I should be worried about? Has anyone else seen this behavior?
Re: 2.5.0 problems on Solaris 8
On Fri, Apr 21, 2006 at 09:41:54AM -0400, stan wrote: I'm starting to upgrade my clients. The HP-UX boxes (10.20) went flawlessly. But I'm having a few issues with my first Solaris 8 machine. Now in the interest of full disclosure, these machines are supported by a 3rd party OEM, and the machine I use for a build machine was recently rebuilt by them to address a kernel memory leak issue, so odds are good what I'm seeing is a result of their meddling, but I could use some help figuring this out. [snip] My first amcheck failed. I tried to run amandad by hand, and found that the dynamic linker was failing to find libgcc_s.so.1. I looked in .cshrc and .profile, and found that they had broken LD_LIBRARY_PATH. I fixed this, restarted inetd, and was able to run an amcheck. So far so good. LD_LIBRARY_PATH is at best a kludge. Needing to use it is always a sign that something's broken. (Having a 537 character PATH generally is, as well.) But, the next backups from this machine failed, and when I look in the debug file, I find this: ld.so.1: /opt/amanda/libexec/noop: fatal: libgcc_s.so.1: open failed: No such file or directory I'm a bit confused by what sets up the search path for shared libraries under Solaris 8, when they are invoked as services like Amanda does. Is there a way to give the binaries a hard coded search path? Yes, pass the -R/path/to/library flag to the linker at build time. For Solaris systems, every -L given to the linker should have a corresponding -R. libgcc_s.so.1 is a bit of a special case, since gcc is implicitly linking that one in when needed. Figure out where it lives and pass the appropriate -R option via LDFLAGS when building.
Re: 2.5.0 problems on Solaris 8
On Fri, Apr 21, 2006 at 11:25:02AM -0700, Mike Delaney wrote: On Fri, Apr 21, 2006 at 09:41:54AM -0400, stan wrote: I'm starting to upgrade my clients. The HP-UX boxes (10.20) went flawlessly. But I'm having a few issues with my first Solaris 8 machine. Now in the interest of full disclosure, these machines are supported by a 3rd party OEM, and the machine I use for a build machine was recently rebuilt by them to address a kernel memory leak issue, so odds are good what I'm seeing is a result of their meddling, but I could use some help figuring this out. [snip] My first amcheck failed. I tried to run amandad by hand, and found that the dynamic linker was failing to find libgcc_s.so.1. I looked in .cshrc and .profile, and found that they had broken LD_LIBRARY_PATH. I fixed this, restarted inetd, and was able to run an amcheck. So far so good. LD_LIBRARY_PATH is at best a kludge. Needing to use it is always a sign that something's broken. (Having a 537 character PATH generally is, as well.) But, the next backups from this machine failed, and when I look in the debug file, I find this: ld.so.1: /opt/amanda/libexec/noop: fatal: libgcc_s.so.1: open failed: No such file or directory I'm a bit confused by what sets up the search path for shared libraries under Solaris 8, when they are invoked as services like Amanda does. Is there a way to give the binaries a hard coded search path? Yes, pass the -R/path/to/library flag to the linker at build time. For Solaris systems, every -L given to the linker should have a corresponding -R. libgcc_s.so.1 is a bit of a special case, since gcc is implicitly linking that one in when needed. Figure out where it lives and pass the appropriate -R option via LDFLAGS when building. OK, should I manual add this to te Makefiles? Or is there a way to tell configure this? Perhaps an environment variable? -- U.S. Encouraged by Vietnam Vote - Officials Cite 83% Turnout Despite Vietcong Terror - New York Times 9/3/1967
Re: 2.5.0 problems on Solaris 8
On Fri, Apr 21, 2006 at 03:47:29PM -0400, stan wrote: On Fri, Apr 21, 2006 at 11:25:02AM -0700, Mike Delaney wrote: Yes, pass the -R/path/to/library flag to the linker at build time. For Solaris systems, every -L given to the linker should have a corresponding -R. libgcc_s.so.1 is a bit of a special case, since gcc is implicitly linking that one in when needed. Figure out where it lives and pass the appropriate -R option via LDFLAGS when building. OK, should I manual add this to te Makefiles? Or is there a way to tell configure this? Perhaps an environment variable? LDFLAGS is an environment variable: $ export LDFLAGS=-R/path/to/gcc/libraries $ ./configure $ make
Re: Dumps performed, but also MISSING ???
Hi,
Could you try this patch?
It's the datestamp in the chunker line that are not correct, they should
be the same as the dumper line.
Jean-Louis
Jean-Louis Martineau wrote:
Hi,
I found the problem, it will report a MISSING for all dump that
doesn't start the same day you run your amdump.
I will provide a patch soon.
Jean-Louis
jeffrey d anderson wrote:
The strange behavior described here started when I upgraded my server
to amanda-2.5.0. I was wondering if anyone else had seen this behavior:
I run amanda 2.5.0 with a Redhat Linux 9 server attached to an ADIC
Scalar 24 changer.
My configuration backs up about 20 clients (mix of linux and SunOS,
running various amanda versions from 2.4.3 to 2.5.0) with about 65
DLEs. I backup about 200GB per night with a dumpcycle of 7 days.
This configuration has been successfully running for about a year,
and seems to still be successful.
So here is the point of this message. I upgraded the server from
amanda 2.4.4p4 to 2.5.0 on March 30 and things appear to be running
fine. Backups proceed and I've done successful restores.
But In the amanda mail report I get messages like this:
DUMP SUMMARY:
DUMPER
STATS TAPER STATS HOSTNAME DISKL
ORIG-kB OUT-kB COMP% client1.here.org / 1
23290 5568 23.9client1.here.org / MISSING
--
client2.here.org -/coldprobe 0 93053120 93053152--
client2.here.org -/coldprobe MISSING-
It looks like these dumps have succeeded, but they are ALSO reported
as missing! Not all DLEs are reported missing, and the MISSING
ones are different each night (different hosts, different DLEs).
Some nights there are no MISSING entries, some nights there are a
couple, some nights there are five or six.
This started with the upgrade to amanda-2.5.0 on the server. No
changes were made to the disklist at that time. I thought perhaps I
had duplicate entries in the disklist file, and amanda's handling of
them had changed, but that does not appear to be the case.
Is this something I should be worried about? Has anyone else seen
this behavior?
diff -u -r --show-c-function --exclude-from=amanda.diff amanda-2.5.0.orig/server-src/chunker.c amanda-2.5.0.new/server-src/chunker.c
--- amanda-2.5.0.orig/server-src/chunker.c 2006-03-28 16:10:06.0 -0500
+++ amanda-2.5.0.new/server-src/chunker.c 2006-04-21 15:57:28.0 -0400
@@ -84,7 +84,7 @@ static char *options = NULL;
static char *progname = NULL;
static int level;
static char *dumpdate = NULL;
-static char *datestamp;
+static char *datestamp = NULL;
static int command_in_transit;
static dumpfile_t file;
@@ -167,7 +167,15 @@ main(main_argc, main_argv)
signal(SIGPIPE, SIG_IGN);
signal(SIGCHLD, SIG_IGN);
-datestamp = construct_datestamp(NULL);
+cmd = getcmd(cmdargs);
+if(cmd == START) {
+ if(cmdargs.argc = 1)
+ error(error [dumper START: not enough args: datestamp]);
+ datestamp = newstralloc(datestamp, cmdargs.argv[2]);
+}
+else {
+ error(Didn't get START command);
+}
/*do {*/
cmd = getcmd(cmdargs);
diff -u -r --show-c-function --exclude-from=amanda.diff amanda-2.5.0.orig/server-src/driver.c amanda-2.5.0.new/server-src/driver.c
--- amanda-2.5.0.orig/server-src/driver.c 2006-04-05 08:27:56.0 -0400
+++ amanda-2.5.0.new/server-src/driver.c 2006-04-21 15:58:22.0 -0400
@@ -791,6 +791,7 @@ start_some_dumps(rq)
chunker-result = LAST_TOK;
dumper-result = LAST_TOK;
startup_chunk_process(chunker,chunker_program);
+ chunker_cmd(chunker, START, (void *)datestamp);
chunker-dumper = dumper;
chunker_cmd(chunker, PORT_WRITE, diskp);
cmd = getresult(chunker-fd, 1, result_argc, result_argv, MAX_ARGS+1);
diff -u -r --show-c-function --exclude-from=amanda.diff amanda-2.5.0.orig/server-src/driverio.c amanda-2.5.0.new/server-src/driverio.c
--- amanda-2.5.0.orig/server-src/driverio.c 2006-03-28 16:10:06.0 -0500
+++ amanda-2.5.0.new/server-src/driverio.c 2006-04-21 16:10:41.0 -0400
@@ -440,12 +440,15 @@ disk_t *dp;
assignedhd_t **h=NULL;
char *features;
-if(dp sched(dp) sched(dp)-holdp) {
+if(cmd != START dp sched(dp) sched(dp)-holdp) {
h = sched(dp)-holdp;
activehd = sched(dp)-activehd;
}
switch(cmd) {
+case START:
+ cmdline = vstralloc(cmdstr[cmd], , (char *)dp, \n, NULL);
+ break;
case PORT_WRITE:
if (dp h) {
holdalloc(h[activehd]-disk)-allocated_dumpers++;
diff -u -r --show-c-function --exclude-from=amanda.diff amanda-2.5.0.orig/server-src/server_util.c amanda-2.5.0.new/server-src/server_util.c
--- amanda-2.5.0.orig/server-src/server_util.c 2006-03-28 16:10:07.0 -0500
+++ amanda-2.5.0.new/server-src/server_util.c 2006-04-21 16:11:53.0 -0400
@@ -35,7 +35,7 @@
const char *cmdstr[] = {
Re: 2.5.0 problems on Solaris 8
On Fri, Apr 21, 2006 at 01:14:13PM -0700, Mike Delaney wrote: On Fri, Apr 21, 2006 at 03:47:29PM -0400, stan wrote: On Fri, Apr 21, 2006 at 11:25:02AM -0700, Mike Delaney wrote: Yes, pass the -R/path/to/library flag to the linker at build time. For Solaris systems, every -L given to the linker should have a corresponding -R. libgcc_s.so.1 is a bit of a special case, since gcc is implicitly linking that one in when needed. Figure out where it lives and pass the appropriate -R option via LDFLAGS when building. OK, should I manual add this to te Makefiles? Or is there a way to tell configure this? Perhaps an environment variable? LDFLAGS is an environment variable: $ export LDFLAGS=-R/path/to/gcc/libraries $ ./configure $ make Thanks. BTW turns out that another kind soul showed me that my shared libary did not have the execute bit set, and that seems to ahve been the root of all evil, in this case. Probably the resuult of the vendor copying things over from the old machine to the new one, incoreclty. Thanks to all for the ehlp. -- U.S. Encouraged by Vietnam Vote - Officials Cite 83% Turnout Despite Vietcong Terror - New York Times 9/3/1967
