Re: [Bacula-users] issue with setuid/gid on restored files
Redhat 6.5 x86_64 On 7/23/14 12:50 AM, Kern Sibbald wrote: Different Linux OSes have very different behaviors, which OS are you running (distribution and version)? On 07/23/2014 12:10 AM, Stephen Thompson wrote: I'm running 7.0.4. Here's an example... (before backup) # ls -ld /bin dr-xr-xr-x 2 root root 4096 Jul 22 09:56 /bin # ls -l /bin/ping -rwsr-xr-x 1 root root 40760 Sep 17 2013 /bin/ping (after restore selecting file /bin/ping) # ls -ld /bin drwsr-xr-x 2 root root 4096 Jul 22 14:38 bin # ls -l /bin/ping -rwxr-xr-x 1 root root 40760 Sep 17 2013 ping (after restore selecting file /bin/ping and directory /bin) # ls -ld /bin dr-xr-xr-x 2 root root 4096 Jul 22 14:38 bin # ls -l /bin/ping -rwxr-xr-x 1 root root 40760 Sep 17 2013 ping In the first restore case, looks like the dir has user-write permissions as well, which isn't right, but perhaps that comes from the umask of the restore since the directory wasn't part of the restore selection. However, the setuid bit certainly wouldn't be coming from the umask. I'm jumping to the conclusion that whatever's doing the setuid bit is messing up and doing it to the parent directory instead of to the file. Stephen On 7/22/14 2:58 PM, Stephen Thompson wrote: Sorry if I have not researched this enough before bringing it to the list, but what I'm seeing is very odd. Someone else must have run into this before me. If I restore a setuid or setgid file, the file is restored without the setuid/setgid bit set. However, the directory containing the file (which did not have it's setuid/setgid bit set during the backup) winds up with the setuid/setgid bit being set. If I restore both the directory and the file, the directory ends up with the proper non-setuid/setgid attributes, but the file once again ends up without the setuid/setgid bit set. I'm assuming the directory has the bit set during an interim stage of the restore, but is then properly set when it's attributes are set during the restore (which must happen after the files that it contains). I can't say authoritatively, but I don't believe this is the way bacula used to behave for me. And to say the least, this is far from acceptable. I discovered this during a bare metal restore, and have loads of issues from no setuid or setgid bits being set on the restored system. thanks, Stephen -- Stephen Thompson Berkeley Seismological Laboratory step...@seismo.berkeley.edu215 McCone Hall # 4760 510.214.6506 (phone) University of California, Berkeley 510.643.5811 (fax) Berkeley, CA 94720-4760 -- Want fast and easy access to all the code in your enterprise? Index and search up to 200,000 lines of code with a free copy of Black Duck Code Sight - the same software that powers the world's largest code search on Ohloh, the Black Duck Open Hub! Try it now. http://p.sf.net/sfu/bds ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] issue with setuid/gid on restored files
compiled from scratch. On 7/23/14 8:02 AM, Simone Caronni wrote: On 23 July 2014 16:18, Kern Sibbald k...@sibbald.com mailto:k...@sibbald.com wrote: On 07/23/2014 04:04 PM, Stephen Thompson wrote: Redhat 6.5 x86_64 OK, that is a particularly tricky system as they have added additional system security which does not permit certain sequences of API calls even as root which other Linux OSes permit :-( I.e. we test on the latest debian/ubuntu and the code works, but not on RHEL 6.x ... I will look at the code as I may have a patch that will help, but I don't remember it having to do with the setuid bit. I recommend that you submit a bug report on this, because if I get distracted this weekend, I might miss coming back to this problem. With a bug report, it remains very visible until it is corrected. Stephen, can you please add me in CC to the bug? I'm the current Fedora Bacula maintainer. BTW, have you compiled Bacula from scratch or used backported packages [1]? Thanks, --Simone [1] http://repos.fedorapeople.org/repos/slaanesh/bacula7/ -- You cannot discover new oceans unless you have the courage to lose sight of the shore (R. W. Emerson). http://xkcd.com/229/ http://negativo17.org/ -- Stephen Thompson Berkeley Seismological Laboratory step...@seismo.berkeley.edu215 McCone Hall # 4760 510.214.6506 (phone) University of California, Berkeley 510.643.5811 (fax) Berkeley, CA 94720-4760 -- Want fast and easy access to all the code in your enterprise? Index and search up to 200,000 lines of code with a free copy of Black Duck Code Sight - the same software that powers the world's largest code search on Ohloh, the Black Duck Open Hub! Try it now. http://p.sf.net/sfu/bds ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
[Bacula-users] 7.0.4 director crashes
Hello, We've had 3 director crashes since updating to 7.0.4, which is highly unusual for us. We've had a stable bacula for years now. Don't know if anyone else has had this issue. We're running on Redhat 6.5 x86_64. I have yet to get a trace. First crash, I hadn't enabled sudo, and the 2-3 crashes, I hadn't disabled sudo's requirement for a tty, so in all three cases btraceback was not able to run properly. I believe I have this resolved in case it crashes again, but I thought I'd ping this list to see if anyone had thoughts. thanks, Stephen -- Stephen Thompson Berkeley Seismological Laboratory step...@seismo.berkeley.edu215 McCone Hall # 4760 510.214.6506 (phone) University of California, Berkeley 510.643.5811 (fax) Berkeley, CA 94720-4760 -- ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] 7.0.4 director crashes
Thanks for the feedback. We've been running on 7.0.4 since June 10th and have had 3 crashes. Have 130+ clients with nightly incrementals and monthly fulls. Stephen On 8/12/14 10:07 AM, Francisco Rafael wrote: I'm using 7.0.5 with 40+ clients, no crash so far... CentOS 6.5 x64. 2014-08-12 13:50 GMT-03:00 John Drescher dresche...@gmail.com mailto:dresche...@gmail.com: We've had 3 director crashes since updating to 7.0.4, which is highly unusual for us. We've had a stable bacula for years now. Don't know if anyone else has had this issue. I have not had any crashes on gentoo with 7.0.4 and 35+ clients. John -- ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net mailto:Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users -- ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users -- Stephen Thompson Berkeley Seismological Laboratory step...@seismo.berkeley.edu215 McCone Hall # 4760 510.214.6506 (phone) University of California, Berkeley 510.643.5811 (fax) Berkeley, CA 94720-4760 -- ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] 7.0.4 director crashes
My most recent crashes created lockdump files, but not my initial one. Stephen On 8/12/14 11:45 AM, Clark, Patricia A. wrote: I have 2 separate instances of Bacula v7.0.5 on RHEL6.5 x86_64. One has had the server FD segfault once. The second instance has had the director segfault twice now. The server has large file systems mounted and is backing these up. It does not have any external clients at this time. It is now generating lockdump files in the spool area when this happens. I have not gone further into debugging as of yet since it has been only on the weekend. Patti Clark Linux System Administrator RD Systems Support Oak Ridge National Laboratory From: Stephen Thompson step...@seismo.berkeley.edumailto:step...@seismo.berkeley.edu Date: Tuesday, August 12, 2014 at 1:20 PM To: bacula-users@lists.sourceforge.netmailto:bacula-users@lists.sourceforge.net bacula-users@lists.sourceforge.netmailto:bacula-users@lists.sourceforge.net Subject: Re: [Bacula-users] 7.0.4 director crashes Thanks for the feedback. We've been running on 7.0.4 since June 10th and have had 3 crashes. Have 130+ clients with nightly incrementals and monthly fulls. Stephen On 8/12/14 10:07 AM, Francisco Rafael wrote: I'm using 7.0.5 with 40+ clients, no crash so far... CentOS 6.5 x64. 2014-08-12 13:50 GMT-03:00 John Drescher dresche...@gmail.commailto:dresche...@gmail.com mailto:dresche...@gmail.com: We've had 3 director crashes since updating to 7.0.4, which is highly unusual for us. We've had a stable bacula for years now. Don't know if anyone else has had this issue. I have not had any crashes on gentoo with 7.0.4 and 35+ clients. John -- ___ Bacula-users mailing list Bacula-users@lists.sourceforge.netmailto:Bacula-users@lists.sourceforge.net mailto:Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users -- ___ Bacula-users mailing list Bacula-users@lists.sourceforge.netmailto:Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users -- Stephen Thompson Berkeley Seismological Laboratory step...@seismo.berkeley.edumailto:step...@seismo.berkeley.edu215 McCone Hall # 4760 510.214.6506 (phone) University of California, Berkeley 510.643.5811 (fax) Berkeley, CA 94720-4760 -- ___ Bacula-users mailing list Bacula-users@lists.sourceforge.netmailto:Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users -- ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users -- Stephen Thompson Berkeley Seismological Laboratory step...@seismo.berkeley.edu215 McCone Hall # 4760 510.214.6506 (phone) University of California, Berkeley 510.643.5811 (fax) Berkeley, CA 94720-4760 -- ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] 7.0.4 director crashes
Additionally, I see that my 'crashes' were segmentation violations. Aug 6 02:20:04 HOST bacula-dir: Bacula interrupted by signal 11: Segmentation violation Aug 7 03:40:05 HOST bacula-dir: Bacula interrupted by signal 11: Segmentation violation Stephen On 8/12/14 1:00 PM, Stephen Thompson wrote: My most recent crashes created lockdump files, but not my initial one. Stephen On 8/12/14 11:45 AM, Clark, Patricia A. wrote: I have 2 separate instances of Bacula v7.0.5 on RHEL6.5 x86_64. One has had the server FD segfault once. The second instance has had the director segfault twice now. The server has large file systems mounted and is backing these up. It does not have any external clients at this time. It is now generating lockdump files in the spool area when this happens. I have not gone further into debugging as of yet since it has been only on the weekend. Patti Clark Linux System Administrator RD Systems Support Oak Ridge National Laboratory From: Stephen Thompson step...@seismo.berkeley.edumailto:step...@seismo.berkeley.edu Date: Tuesday, August 12, 2014 at 1:20 PM To: bacula-users@lists.sourceforge.netmailto:bacula-users@lists.sourceforge.net bacula-users@lists.sourceforge.netmailto:bacula-users@lists.sourceforge.net Subject: Re: [Bacula-users] 7.0.4 director crashes Thanks for the feedback. We've been running on 7.0.4 since June 10th and have had 3 crashes. Have 130+ clients with nightly incrementals and monthly fulls. Stephen On 8/12/14 10:07 AM, Francisco Rafael wrote: I'm using 7.0.5 with 40+ clients, no crash so far... CentOS 6.5 x64. 2014-08-12 13:50 GMT-03:00 John Drescher dresche...@gmail.commailto:dresche...@gmail.com mailto:dresche...@gmail.com: We've had 3 director crashes since updating to 7.0.4, which is highly unusual for us. We've had a stable bacula for years now. Don't know if anyone else has had this issue. I have not had any crashes on gentoo with 7.0.4 and 35+ clients. John -- ___ Bacula-users mailing list Bacula-users@lists.sourceforge.netmailto:Bacula-users@lists.sourceforge.net mailto:Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users -- ___ Bacula-users mailing list Bacula-users@lists.sourceforge.netmailto:Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users -- Stephen Thompson Berkeley Seismological Laboratory step...@seismo.berkeley.edumailto:step...@seismo.berkeley.edu215 McCone Hall # 4760 510.214.6506 (phone) University of California, Berkeley 510.643.5811 (fax) Berkeley, CA 94720-4760 -- ___ Bacula-users mailing list Bacula-users@lists.sourceforge.netmailto:Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users -- ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users -- Stephen Thompson Berkeley Seismological Laboratory step...@seismo.berkeley.edu215 McCone Hall # 4760 510.214.6506 (phone) University of California, Berkeley 510.643.5811 (fax) Berkeley, CA 94720-4760 -- ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
[Bacula-users] solaris sparc 7.0.5 clients crash
Anyone with success in running a 7x client on Solaris 10 SPARC? We've recently attempted to upgrade clients from 5x to 7.0.5 and it works fine on Solaris 10 x86, but on SPARC nothing but crashes once jobs are submitted. SPARC clients build and run (without jobs) fine. http://bugs.bacula.org/view.php?id=2094 thanks, Stephen -- Stephen Thompson Berkeley Seismological Laboratory step...@seismo.berkeley.edu215 McCone Hall # 4760 510.214.6506 (phone) University of California, Berkeley 510.643.5811 (fax) Berkeley, CA 94720-4760 -- ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] solaris sparc 7.0.5 clients crash
Additionally, we run with Accurate backups. It looks like the crash may be occurring between the time the SD sends the list for accurate backups but before the client traverses the fileset. Stephen On 8/19/14 7:47 AM, Stephen Thompson wrote: Anyone with success in running a 7x client on Solaris 10 SPARC? We've recently attempted to upgrade clients from 5x to 7.0.5 and it works fine on Solaris 10 x86, but on SPARC nothing but crashes once jobs are submitted. SPARC clients build and run (without jobs) fine. http://bugs.bacula.org/view.php?id=2094 thanks, Stephen -- Stephen Thompson Berkeley Seismological Laboratory step...@seismo.berkeley.edu215 McCone Hall # 4760 510.214.6506 (phone) University of California, Berkeley 510.643.5811 (fax) Berkeley, CA 94720-4760 -- ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] solaris sparc 7.0.5 clients crash
--tag=CXX --mode=link /usr/sfw/bin/g++ -shared bpipe-fd.lo -o bpipe-fd.la -rpath /opt/bacula/lib -module -export-dynamic -avoid-version /opt/src/bacula/bacula-7.0.5-CLIENT/libtool --silent --tag=CXX --mode=compile /usr/sfw/bin/g++ -fno-strict-aliasing -fno-exceptions -fno-rtti -g -O2 -Wall -fno-strict-aliasing -fno-exceptions -fno-rtti -I../.. -I../../filed -c test-plugin-fd.c /opt/src/bacula/bacula-7.0.5-CLIENT/libtool --silent --tag=CXX --mode=link /usr/sfw/bin/g++ -shared test-plugin-fd.lo -o test-plugin-fd.la -rpath /opt/bacula/lib -module -export-dynamic -avoid-version /opt/src/bacula/bacula-7.0.5-CLIENT/libtool --silent --tag=CXX --mode=compile /usr/sfw/bin/g++ -fno-strict-aliasing -fno-exceptions -fno-rtti -g -O2 -Wall -fno-strict-aliasing -fno-exceptions -fno-rtti -I../.. -I../../filed -c test-deltaseq-fd.c /opt/src/bacula/bacula-7.0.5-CLIENT/libtool --silent --tag=CXX --mode=link /usr/sfw/bin/g++ -shared test-deltaseq-fd.lo -o test-deltaseq-fd.la -rpath /opt/bacula/lib -module -export-dynamic -avoid-version ==Entering directory /opt/src/bacula/bacula-7.0.5-CLIENT/manpages On 8/19/14 8:31 AM, Heitor Faria wrote: Stephen, Sorry for insisting on this subject, but I saw that even you using the --enable-client-only the configuration output said it would build Director: client-only:yes build-dird: yes build-stored: yes Last night I've compiled for Debian, and if the MYSQL_LIBS path wasn't correct the director and file daemon were not built. Again: this is just a wild guess that could be tested. Sorry I don't have a Solaris here installed to test for you. It's a client-only build not linked against any database. env CC='/usr/sfw/bin/gcc' \ env CXX='/usr/sfw/bin/g++' \ env CFLAGS='-g -O2' \ env CXXFLAGS='-g -02' \ ./configure \ --prefix=$BHOME \ --sbindir=$BHOME/bin \ --sysconfdir=$BHOME/conf \ --with-working-dir=$BHOME/work \ --with-bsrdir=$BHOME/log \ --with-logdir=$BHOME/log \ --with-pid-dir=/var/run \ --with-subsys-dir=/var/run \ --with-basename=lawson \ --with-hostname=lawson \ --with-dump-email=$EMAIL \ --enable-smartalloc \ --enable-client-only \ --with-openssl=no Same configure works fine with 5X source; been running with this for literally years, though many versions of 5X. Same config works fine on Solaris 10 x86. Stephen On 8/19/14 8:08 AM, Heitor Faria wrote: Anyone with success in running a 7x client on Solaris 10 SPARC? We've recently attempted to upgrade clients from 5x to 7.0.5 and it works fine on Solaris 10 x86, but on SPARC nothing but crashes once jobs are submitted. SPARC clients build and run (without jobs) fine. Couldn't autenticate in the link but wild hint here: did you changed the /src/cats/Makefile to put the correct path to the database libs? http://bugs.bacula.org/view.__php?id=2094 http://bugs.bacula.org/view.php?id=2094 thanks, Stephen -- Stephen Thompson Berkeley Seismological Laboratory step...@seismo.berkeley.edu mailto:step...@seismo.berkeley.edu mailto:stephen@seismo.__berkeley.edu mailto:step...@seismo.berkeley.edu 215 McCone Hall # 4760 510.214.6506 tel:510.214.6506 (phone) University of California, Berkeley 510.643.5811 tel:510.643.5811 (fax) Berkeley, CA 94720-4760 --__--__-- _ Bacula-users mailing list Bacula-users@lists.__sourceforge.net mailto:Bacula-users@lists.sourceforge.net mailto:Bacula-users@lists.__sourceforge.net mailto:Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/__lists/listinfo/bacula-users https://lists.sourceforge.net/lists/listinfo/bacula-users -- Stephen Thompson Berkeley Seismological Laboratory step...@seismo.berkeley.edu mailto:step...@seismo.berkeley.edu 215 McCone Hall # 4760 510.214.6506 tel:510.214.6506 (phone) University of California, Berkeley 510.643.5811 tel:510.643.5811 (fax) Berkeley, CA 94720-4760 -- Heitor Medrado de Faria | Need Bacula training? 10% discount coupon code at Udemy: bacula-users https://www.udemy.com/bacula-backup-software/?couponCode=bacula-users +55 61 2021-8260 +55 61 8268-4220 Site: www.bacula.com.br
Re: [Bacula-users] solaris sparc 7.0.5 clients crash
Ah. I think that fixed it. Thanks! On 8/19/14 10:28 AM, Martin Simmons wrote: On Tue, 19 Aug 2014 07:47:39 -0700, Stephen Thompson said: Anyone with success in running a 7x client on Solaris 10 SPARC? We've recently attempted to upgrade clients from 5x to 7.0.5 and it works fine on Solaris 10 x86, but on SPARC nothing but crashes once jobs are submitted. SPARC clients build and run (without jobs) fine. http://bugs.bacula.org/view.php?id=2094 Maybe a compiler bug? Try without -O2. __Martin -- ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users -- Stephen Thompson Berkeley Seismological Laboratory step...@seismo.berkeley.edu215 McCone Hall # 4760 510.214.6506 (phone) University of California, Berkeley 510.643.5811 (fax) Berkeley, CA 94720-4760 -- ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
[Bacula-users] Error: block.c:255 Write errors?
Hello, I sporadically get these types of alerts for one on my bacula tape libraries... 05-Sep 00:41 lawson-sd_L100_ JobId 389348: Error: block.c:255 Write error at 610:412 on device L100-Drive-0 (/dev/L100-Drive-0). ERR=Input/output error. Am I correct in assuming that this was indeed a tape write error, but that bacula will attempt a 2nd write of the same block of data and if that 2nd attempt succeeds proceed on and ultimately have a successfully run job (one that can be restored without issue)? In other words, should this error worry me if it doesn't happen often? It does consistently happen -- with 100's of jobs a night, it probably happens 3-4 times a week. thanks, Stephen -- Slashdot TV. Video for Nerds. Stuff that matters. http://tv.slashdot.org/ ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] Error: block.c:255 Write errors?
Huh, maybe this is a misdiagnosis of the end of tape and a write error only in the sense that there is no tape left. 05-Sep 00:41 SD_L100_ JobId 389348: Error: block.c:255 Write error at 610:412 on device L100-Drive-0 (/dev/L100-Drive-0). ERR=Input/output error. 05-Sep 00:41 SD_L100_ JobId 389348: Re-read of last block succeeded. 05-Sep 00:41 SD_L100_ JobId 389348: End of medium on Volume IM0161 Bytes=1,090,307,051,520 Blocks=520,103 at 05-Sep-2014 00:41. On 09/05/2014 09:42 AM, Stephen Thompson wrote: Hello, I sporadically get these types of alerts for one on my bacula tape libraries... 05-Sep 00:41 lawson-sd_L100_ JobId 389348: Error: block.c:255 Write error at 610:412 on device L100-Drive-0 (/dev/L100-Drive-0). ERR=Input/output error. Am I correct in assuming that this was indeed a tape write error, but that bacula will attempt a 2nd write of the same block of data and if that 2nd attempt succeeds proceed on and ultimately have a successfully run job (one that can be restored without issue)? In other words, should this error worry me if it doesn't happen often? It does consistently happen -- with 100's of jobs a night, it probably happens 3-4 times a week. thanks, Stephen -- Slashdot TV. Video for Nerds. Stuff that matters. http://tv.slashdot.org/ ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users -- Slashdot TV. Video for Nerds. Stuff that matters. http://tv.slashdot.org/ ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] 7.2 mysql issue?
Yes, we've tuned the database a number of times and believe it's the best we can do. On 10/14/15 3:17 AM, Alex Domoradov wrote: > The same thing as for me. I'm trying do not use mysql shipped with > CentOS 6 and replace it with Percona 5.5/5.6 whenever it's possible > > 2 Stephen > Have you tried to run mysqltunner? > > > On Mon, Oct 12, 2015 at 07:33:46AM -0700, Stephen Thompson wrote: > > > > update... > > > > After adding more RAM, we are back to getting a about 3 queries a day > > that run longer than 15 minutes. This was our norm before upgrading. > > No job errors since the first couple days from this month (Oct). Not > > sure if the reduction in long running queries was actually from > > additional RAM or not, since last week before adding RAM, the number of > > long running queries per day had already greatly diminished since > > beginning of month. > > > > So, I guess, problem solved for now, though I'm not completely confident > > about what actually happened or if I did anything to fix it. > > Oh, well. > > > > Stephen > > Hi Stephen, > > you might also try giving MariaDB a shot which has been performing > fine as a drop-in mysql replacement for us for the last few years with > catalogs of similar size. > > Cheers, Uwe > > > > > > > > > > > -- > ___ > Bacula-users mailing list > Bacula-users@lists.sourceforge.net > <mailto:Bacula-users@lists.sourceforge.net> > https://lists.sourceforge.net/lists/listinfo/bacula-users > > -- Stephen Thompson Berkeley Seismological Laboratory step...@seismo.berkeley.edu215 McCone Hall #4760 Office: 510.664.9177 University of California, Berkeley Remote: 510.214.6506 (Tue,Wed) Berkeley, CA 94720-4760 -- ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] 7.2 mysql issue?
update... After adding more RAM, we are back to getting a about 3 queries a day that run longer than 15 minutes. This was our norm before upgrading. No job errors since the first couple days from this month (Oct). Not sure if the reduction in long running queries was actually from additional RAM or not, since last week before adding RAM, the number of long running queries per day had already greatly diminished since beginning of month. So, I guess, problem solved for now, though I'm not completely confident about what actually happened or if I did anything to fix it. Oh, well. Stephen On 10/9/15 2:08 PM, Stephen Thompson wrote: > > > Eric, > > I appreciate all the feedback. We went through a few iterations of > tuning awhile back and have not generally had any significant issues > over the years with database responsiveness. > > Back to the original post, it's only been since our upgrade that we > started having database lock timeout issues. Otherwise we've run for > years (6 or so) without issue. We also went through an orphan record > cleanout earlier this year. > > Stat wise, it looks like our slow queries are still happening at twice > the rate compared to recent months, but half as often as they were when > I first reported the issue a week ago, so I am equally nonplussed about > the improvement as I was about the lockouts. > > I did get a chance to double the ram from 8 to 16GB today though > unfortunately we don't have the ready resources to do many hardware > upgrades, though I quite understand why that's a recommendation. > > Stephen > > > > On 10/08/2015 10:58 PM, Eric Bollengier wrote: >> Hello Stephen, >> >> >> Le 05. 10. 15 19:17, Stephen Thompson a écrit : >>> >>> Eric, >>> >>> Thanks for the reply. >>> >>> I've heard the postgres recommendation a fair number of times. A couple >>> years back, we setup a parallel instance but even after tuning still >>> wound up with _worse_ performance than with mysql. I could not figure >>> out what to attribute this to (because it was in such contrast to all >>> the pro-postgres recommendations) except possibly our memory-poor server >>> - 8Gb RAM. >>> >>> At any rate, the only thing that's changed was the upgrade from 7.0.5 to >>> 7.2.0. The table involved is definitely the File table. We do have >>> jobs with 20-30 million records, so those jobs can be slow when it comes >>> time for attribute insertion into the database (or to read out a file >>> list for Accurate backups). This why we've historically had innodb lock >>> timeout of 3600. However, it's only last week after the upgrade that >>> we've ever had queries extend beyond that hour mark. >>> >>> We also went through a database cleaning process last month due to >>> nearly reaching 1Tb and I can pretty authoritatively claim that we don't >>> have orphan records. The database content and schema all appear to be >>> appropriate. >> >> A 1TB database (running either Postgresql, MySQL or whatever other kind >> of product) should be carefully tuned and monitored. My guess would be >> that your my.cnf settings are not suitable for such database size. You >> can run a tool such as MySQLtuner to check that everything is ok on >> MySQL side, increase the size of the memory of your server or try to >> cleanup orphan filename records. >> >> The size of the File table should not impact performances on Backup, but >> other tables such as Path or Filename are important (and they are pretty >> big on your site). >> >> > I was worried that queries had been rewritten that made it >> > more efficient for other databases, but less so for mysql. >> >> We didn't wrote database query specifically for PostgreSQL or MySQL but >> we optimize them when it's possible, some SQLite queries were optimized >> by a contributor 2 or 3 years ago, and it was way faster for some parts >> of Bacula afterward. >> >> If you look the database world from outside, you might think that >> everything is nice and smooth because all products seem to talk the >> same language (SQL), but they all have a different way to handle the >> work and the SQL specifications (and the lack of specifications). >> For myself, I'm a PostgreSQL user for a quite long time, I have good >> relationships with the PostgreSQL community, and we got huge help when >> we wrote the "Batch Mode" few years ago. I know that it works well and >> we can analyze problems quite easily, doing so I always advise strongly >> to use PostgreSQ
Re: [Bacula-users] 7.2 mysql issue?
mysql> show indexes from File; +---++--+--+-+---+-+--++--++-+ | Table | Non_unique | Key_name | Seq_in_index | Column_name | Collation | Cardinality | Sub_part | Packed | Null | Index_type | Comment | +---++--+--+-+---+-+--++--++-+ | File | 0 | PRIMARY |1 | FileId | A | 4494348205 | NULL | NULL | | BTREE | | | File | 1 | JobId|1 | JobId | A | 19 | NULL | NULL | | BTREE | | | File | 1 | JobId|2 | PathId | A | 408577109 | NULL | NULL | | BTREE | | | File | 1 | JobId|3 | FilenameId | A | 4494348205 | NULL | NULL | | BTREE | | +---++--+--+-+---+-+--++--++-+ On 10/05/2015 10:30 AM, Stephen Thompson wrote: > > > Phil, > > Good question. I vaguely recollect doing that a few years back, but I > don't immediately see any additional indexing. Where can I reference > what the default indexes are supposed to be? > > thanks, > Stephen > > > > On 10/05/2015 10:28 AM, Phil Stracchino wrote: >> On 10/05/15 13:17, Stephen Thompson wrote: >>> At any rate, the only thing that's changed was the upgrade from 7.0.5 to >>> 7.2.0. The table involved is definitely the File table. We do have >>> jobs with 20-30 million records, so those jobs can be slow when it comes >>> time for attribute insertion into the database (or to read out a file >>> list for Accurate backups). This why we've historically had innodb lock >>> timeout of 3600. However, it's only last week after the upgrade that >>> we've ever had queries extend beyond that hour mark. >> >> Stephen, >> Just as a thought, there have been a number of threads on this mailing >> list recommending additional or modified indexes on the File table. >> Have you added the suggested additional indexes? >> >> > -- Stephen Thompson Berkeley Seismological Laboratory step...@seismo.berkeley.edu215 McCone Hall # 4760 Office: 510.664.9177 University of California, Berkeley Remote: 510.214.6506 (Tue,Wed) Berkeley, CA 94720-4760 -- ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] 7.2 mysql issue?
Nevermind about question concerning Snapshot table. I see what happened there. On 10/05/2015 10:17 AM, Stephen Thompson wrote: > > Eric, > > Thanks for the reply. > > I've heard the postgres recommendation a fair number of times. A couple > years back, we setup a parallel instance but even after tuning still > wound up with _worse_ performance than with mysql. I could not figure > out what to attribute this to (because it was in such contrast to all > the pro-postgres recommendations) except possibly our memory-poor server > - 8Gb RAM. > > At any rate, the only thing that's changed was the upgrade from 7.0.5 to > 7.2.0. The table involved is definitely the File table. We do have > jobs with 20-30 million records, so those jobs can be slow when it comes > time for attribute insertion into the database (or to read out a file > list for Accurate backups). This why we've historically had innodb lock > timeout of 3600. However, it's only last week after the upgrade that > we've ever had queries extend beyond that hour mark. > > We also went through a database cleaning process last month due to > nearly reaching 1Tb and I can pretty authoritatively claim that we don't > have orphan records. The database content and schema all appear to be > appropriate. I was worried that queries had been rewritten that made it > more efficient for other databases, but less so for mysql. > > > More info... > > example from slow query logfile: > # Time: 151001 1:28:14 > # User@Host: bacula[bacula] @ localhost [] > # Query_time: 3675.052083 Lock_time: 73.719795 Rows_sent: 0 > Rows_examined: 3 > SET timestamp=1443688094; > INSERT INTO File (FileIndex, JobId, PathId, FilenameId, LStat, MD5, > DeltaSeq) SELECT batch.FileIndex, batch.JobId, Path.PathId, > Filename.FilenameId,batch.LStat, batch.MD5, batch.DeltaSeq FROM batch > JOIN Path ON (batch.Path = Path.Path) JOIN Filename ON (batch.Name = > Filename.Name); > > mysqld: > mysql-5.1.73-5.el6_6.x86_64 > > record counts per table: > File 4,315,675,600 > Filename 154,748,787 > Path 28,534,411 > > innodb file sizes: > 847708500 File.ibd > 19488772 Filename.ibd > 8216580 Path.ibd > 106500PathHierarchy.ibd > 57344 JobMedia.ibd > 40960 PathVisibility.ibd > 27648 Job.ibd > 512 Media.ibd > 176 FileSet.ibd > 144 JobHisto.ibd > 144 Client.ibd > 112 RestoreObject.ibd > 112 Pool.ibd > 112 Log.ibd > 112 BaseFiles.ibd > 96Version.ibd > 96UnsavedFiles.ibd > 96Storage.ibd > 96Status.ibd > 96MediaType.ibd > 96LocationLog.ibd > 96Location.ibd > 96Device.ibd > 96Counters.ibd > 96CDImages.ibd > 4 Snapshot.MYI > 0 Snapshot.MYD > > > > Not related, but I just noticed that somehow the new Snapshot table is > MyISAM format. How did that happen? > > Regarding: > > Would be nice also if you can give the number of Filename per Client > (from the job table). > > Do you have a sample SQL to retrieve this stat? > > > thanks, > Stephen > > > > > > > > On 10/03/2015 12:02 AM, Eric Bollengier wrote: >> Hello Stephen, >> >> On 10/03/2015 12:00 AM, Stephen Thompson wrote: >>> >>> >>> All, >>> >>> I believe I'm having mysql database issues since upgrading to 7.2 (from >>> 7.0.2). I run mysql innodb with 900Gb database that's largely the File >>> table. >> >> For large catalog, we usually advise to use PostgreSQL where we have >> multi-terabytes databases in production. >> >>> Since upgrading, I lose a few jobs a night due to database locking >>> timeouts, which I have set to 3600. I also log slow queries. >> >> Can you get some information about these locks? On which table? Can you >> give some statistics on your catalog like the size and the number of >> records of the File, Filename and Path table? Would be nice also if you >> can give the number of Filename per Client (from the job table). >> >> You might have many orphan Filenames, and MySQL is not always very good >> to join large tables (it uses nested loops, and cannot use the index on >> the Text column in all queries). >> >>> It appears that typically during a months I have about 90-100 queries >>> that take longer than 15 minutes to run. Already this month (upgraded >>> earlier this week), I have 32 queries that take longer than 15 minutes.
Re: [Bacula-users] 7.2 mysql issue?
Phil, Good question. I vaguely recollect doing that a few years back, but I don't immediately see any additional indexing. Where can I reference what the default indexes are supposed to be? thanks, Stephen On 10/05/2015 10:28 AM, Phil Stracchino wrote: > On 10/05/15 13:17, Stephen Thompson wrote: >> At any rate, the only thing that's changed was the upgrade from 7.0.5 to >> 7.2.0. The table involved is definitely the File table. We do have >> jobs with 20-30 million records, so those jobs can be slow when it comes >> time for attribute insertion into the database (or to read out a file >> list for Accurate backups). This why we've historically had innodb lock >> timeout of 3600. However, it's only last week after the upgrade that >> we've ever had queries extend beyond that hour mark. > > Stephen, > Just as a thought, there have been a number of threads on this mailing > list recommending additional or modified indexes on the File table. > Have you added the suggested additional indexes? > > -- Stephen Thompson Berkeley Seismological Laboratory step...@seismo.berkeley.edu215 McCone Hall # 4760 Office: 510.664.9177 University of California, Berkeley Remote: 510.214.6506 (Tue,Wed) Berkeley, CA 94720-4760 -- ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] 7.2 mysql issue?
Thanks for the help. Though, this is giving me a syntax error. ERROR 1064 (42000): You have an error in your SQL syntax; check the manual that corresponds to your MySQL server version for the right syntax to use near 'select Client.Name, count(distinct Filename.FilenameId) from Client, Filen' at line 1 On 10/7/15 6:23 AM, Ana Emília M. Arruda wrote: > select Client.Name, count(distinct Filename.FilenameId) from Client, > Filename, File, Job where Filename.FilenameId=File.FilenameId and > File.JobId=Job.JobId and Job.ClientId=Client.ClientId group by > Client.ClientId; -- Stephen Thompson Berkeley Seismological Laboratory step...@seismo.berkeley.edu215 McCone Hall #4760 Office: 510.664.9177 University of California, Berkeley Remote: 510.214.6506 (Tue,Wed) Berkeley, CA 94720-4760 -- Full-scale, agent-less Infrastructure Monitoring from a single dashboard Integrate with 40+ ManageEngine ITSM Solutions for complete visibility Physical-Virtual-Cloud Infrastructure monitoring from one console Real user monitoring with APM Insights and performance trend reports Learn More http://pubads.g.doubleclick.net/gampad/clk?id=247754911=/4140 ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] 7.2 mysql issue?
Eric, I appreciate all the feedback. We went through a few iterations of tuning awhile back and have not generally had any significant issues over the years with database responsiveness. Back to the original post, it's only been since our upgrade that we started having database lock timeout issues. Otherwise we've run for years (6 or so) without issue. We also went through an orphan record cleanout earlier this year. Stat wise, it looks like our slow queries are still happening at twice the rate compared to recent months, but half as often as they were when I first reported the issue a week ago, so I am equally nonplussed about the improvement as I was about the lockouts. I did get a chance to double the ram from 8 to 16GB today though unfortunately we don't have the ready resources to do many hardware upgrades, though I quite understand why that's a recommendation. Stephen On 10/08/2015 10:58 PM, Eric Bollengier wrote: > Hello Stephen, > > > Le 05. 10. 15 19:17, Stephen Thompson a écrit : >> >> Eric, >> >> Thanks for the reply. >> >> I've heard the postgres recommendation a fair number of times. A couple >> years back, we setup a parallel instance but even after tuning still >> wound up with _worse_ performance than with mysql. I could not figure >> out what to attribute this to (because it was in such contrast to all >> the pro-postgres recommendations) except possibly our memory-poor server >> - 8Gb RAM. >> >> At any rate, the only thing that's changed was the upgrade from 7.0.5 to >> 7.2.0. The table involved is definitely the File table. We do have >> jobs with 20-30 million records, so those jobs can be slow when it comes >> time for attribute insertion into the database (or to read out a file >> list for Accurate backups). This why we've historically had innodb lock >> timeout of 3600. However, it's only last week after the upgrade that >> we've ever had queries extend beyond that hour mark. >> >> We also went through a database cleaning process last month due to >> nearly reaching 1Tb and I can pretty authoritatively claim that we don't >> have orphan records. The database content and schema all appear to be >> appropriate. > > A 1TB database (running either Postgresql, MySQL or whatever other kind > of product) should be carefully tuned and monitored. My guess would be > that your my.cnf settings are not suitable for such database size. You > can run a tool such as MySQLtuner to check that everything is ok on > MySQL side, increase the size of the memory of your server or try to > cleanup orphan filename records. > > The size of the File table should not impact performances on Backup, but > other tables such as Path or Filename are important (and they are pretty > big on your site). > > > I was worried that queries had been rewritten that made it > > more efficient for other databases, but less so for mysql. > > We didn't wrote database query specifically for PostgreSQL or MySQL but > we optimize them when it's possible, some SQLite queries were optimized > by a contributor 2 or 3 years ago, and it was way faster for some parts > of Bacula afterward. > > If you look the database world from outside, you might think that > everything is nice and smooth because all products seem to talk the > same language (SQL), but they all have a different way to handle the > work and the SQL specifications (and the lack of specifications). > For myself, I'm a PostgreSQL user for a quite long time, I have good > relationships with the PostgreSQL community, and we got huge help when > we wrote the "Batch Mode" few years ago. I know that it works well and > we can analyze problems quite easily, doing so I always advise strongly > to use PostgreSQL for all large setup. For other products, developers > uses MySQL and the PostgreSQL driver is not good at all. > > With the time, I found that you can do "more" with "less" hardware when > using the PostgreSQL catalog. In your case (a fairly big database), it > might be the time to spend a bit of money to get more RAM and/or make > sure that your Path/Filename indexes stay in RAM. > > > Hope it helps. > > Best Regards, > Eric > >> >> >> More info... >> >> example from slow query logfile: >> # Time: 151001 1:28:14 >> # User@Host: bacula[bacula] @ localhost [] >> # Query_time: 3675.052083 Lock_time: 73.719795 Rows_sent: 0 >> Rows_examined: 3 >> SET timestamp=1443688094; >> INSERT INTO File (FileIndex, JobId, PathId, FilenameId, LStat, MD5, >> DeltaSeq) SELECT batch.FileIndex, batch.JobId, Path.PathId, >> Filename.FilenameId,batch.LStat, batc
Re: [Bacula-users] 7.2 mysql issue?
973 | | 41 | 219826 | | 53 | 219767 | | 63 | 219749 | | 135 | 219746 | | 141 | 219344 | | 124 | 219157 | | 57 | 219070 | | 134 | 215349 | | 227 | 154642 | | 112 | 134792 | | 125 | 114623 | | 31 | 99493 | | 49 | 98341 | | 34 | 92193 | | 50 | 90190 | | 46 | 88746 | | 111 | 87960 | | 148 | 70591 | | 62 | 68151 | | 145 | 65377 | | 42 | 65290 | | 25 | 63220 | | 60 | 62653 | | 38 | 62183 | | 43 | 46063 | | 228 | 45989 | | 44 | 45433 | | 113 | 44317 | | 186 | 1 | |5 | 0 | | 56 | 0 | | 172 | 0 | | 195 | 0 | | 174 | 0 | | 48 | 0 | | 61 | 0 | +--++ 221 rows in set (0.21 sec) On 10/09/2015 10:01 AM, Eric Bollengier wrote: > Very good point Ana, > > So, you might want to add to the query "AND PurgedFiles = 0" > > Thanks, > > Eric > > Le 09. 10. 15 14:24, Ana Emília M. Arruda a écrit : >> Hello Eric! >> >> Thank you. I thought that you were looking for the number of filename >> per Client that had not been pruned yet :). >> >> Best regards, >> Ana >> >> On Fri, Oct 9, 2015 at 3:17 AM, Eric Bollengier >> <eric.bolleng...@baculasystems.com >> <mailto:eric.bolleng...@baculasystems.com>> wrote: >> >> Thanks Ana! >> >> Something such as >> >> SELECT ClientId, SUM(JobFiles) AS NB FROM Job GROUP BY ClientId >> ORDER BY NB DESC; >> >> should also do the trick a bit more faster ;-) >> >> Best Regards, >> Eric >> >> Le 07. 10. 15 15:23, Ana Emília M. Arruda a écrit : >> >> Hello Stephen, >> >> On Mon, Oct 5, 2015 at 2:17 PM, Stephen Thompson >> <step...@seismo.berkeley.edu >> <mailto:step...@seismo.berkeley.edu> >> <mailto:step...@seismo.berkeley.edu >> <mailto:step...@seismo.berkeley.edu>>> wrote: >> >> >> Regarding: >>> Would be nice also if you can give the number of >> Filename per Client >> (from the job table). >> >> Do you have a sample SQL to retrieve this stat? >> >> >> select Client.Name, count(distinct Filename.FilenameId) from >> Client, >> Filename, File, Job where Filename.FilenameId=File.FilenameId and >> File.JobId=Job.JobId and Job.ClientId=Client.ClientId group by >> Client.ClientId; >> >> The above query should work. >> >> Best regards, >> Ana >> >> >> >> thanks, >> Stephen >> >> >> >> >> >> >> >> On 10/03/2015 12:02 AM, Eric Bollengier wrote: >> > Hello Stephen, >> > >> > On 10/03/2015 12:00 AM, Stephen Thompson wrote: >> >> >> >> >> >> All, >> >> >> >> I believe I'm having mysql database issues since >> upgrading to >> 7.2 (from >> >> 7.0.2). I run mysql innodb with 900Gb database that's >> largely >> the File >> >> table. >> > >> > For large catalog, we usually advise to use PostgreSQL >> where we have >> > multi-terabytes databases in production. >> > >> >> Since upgrading, I lose a few jobs a night due to >> database locking >> >> timeouts, which I have set to 3600. I also log slow >> queries. >> > >> > Can you get some information about these locks? On which >> table? >> Can you >> > give some statistics on your catalog like the size and >> the number of >> > records of the File, Filename and Path table? Would be >> nice also >> if you >> > can give the number of Filename per Client (from the job >> table). >> > >> > Y
[Bacula-users] 7.2 mysql issue?
All, I believe I'm having mysql database issues since upgrading to 7.2 (from 7.0.2). I run mysql innodb with 900Gb database that's largely the File table. Since upgrading, I lose a few jobs a night due to database locking timeouts, which I have set to 3600. I also log slow queries. It appears that typically during a months I have about 90-100 queries that take longer than 15 minutes to run. Already this month (upgraded earlier this week), I have 32 queries that take longer than 15 minutes. At this rate (after 2 days) that will up my regular average of 90-100 to 480! Something is wrong and the coincidence is pretty strong that it's related to the upgrade. Ideas? thanks, Stephen On 09/25/2015 09:02 AM, Stephen Thompson wrote: > > > So far so good. Minor snafu on my part when updating database, but I'm > running 7.2 now. Looking good so far. Will find out more when hundreds > of jobs run tonight. > > Stephen > > > > On 09/24/2015 08:40 AM, Stephen Thompson wrote: >> >> All, >> >> I typically patch bacula pretty frequently, but I saw the somewhat >> unusual notice on the latest release notes that warns it may not be >> ready for use in production. How stable is it? I don't really have the >> resources to test this out, but rather would have to go straight to >> production with it. I could always roll back, but that might entail the >> recovery from dump of a 900GB database. Opinions? >> >> thanks, >> Stephen >> > -- Stephen Thompson Berkeley Seismological Laboratory step...@seismo.berkeley.edu215 McCone Hall # 4760 Office: 510.664.9177 University of California, Berkeley Remote: 510.214.6506 (Tue,Wed) Berkeley, CA 94720-4760 -- ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
[Bacula-users] bacula-fd.service systemd file?
All, I build bacula 7.2 on rhel 7.1 but have not systemd file for bacula-fd. Is there an example available? I thought perhaps that building bacula would make one, as I have this at the end of my configure output: systemd support: yes /etc/systemd/system But I do not appear to see any systemd file example in the source tree. Am I just not looking in the right place? If one does not exist, does anyone have one that I could see? thanks, Stephen -- Stephen Thompson Berkeley Seismological Laboratory step...@seismo.berkeley.edu215 McCone Hall #4760 Office: 510.664.9177 University of California, Berkeley Remote: 510.214.6506 (Tue,Wed) Berkeley, CA 94720-4760 -- Monitor Your Dynamic Infrastructure at Any Scale With Datadog! Get real-time metrics from all of your servers, apps and tools in one place. SourceForge users - Click here to start your Free Trial of Datadog now! http://pubads.g.doubleclick.net/gampad/clk?id=241902991=/4140 ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] bacula-fd.service systemd file?
Sorry, I don't know how I missed this before in src tree... ./platforms/systemd/bacula-fd.service On 9/23/15 10:45 AM, Stephen Thompson wrote: > > All, > > I build bacula 7.2 on rhel 7.1 but have not systemd file for bacula-fd. > Is there an example available? > > I thought perhaps that building bacula would make one, as I have this at > the end of my configure output: > > systemd support: yes /etc/systemd/system > > But I do not appear to see any systemd file example in the source tree. > Am I just not looking in the right place? If one does not exist, does > anyone have one that I could see? > > thanks, > Stephen -- Stephen Thompson Berkeley Seismological Laboratory step...@seismo.berkeley.edu215 McCone Hall #4760 Office: 510.664.9177 University of California, Berkeley Remote: 510.214.6506 (Tue,Wed) Berkeley, CA 94720-4760 -- Monitor Your Dynamic Infrastructure at Any Scale With Datadog! Get real-time metrics from all of your servers, apps and tools in one place. SourceForge users - Click here to start your Free Trial of Datadog now! http://pubads.g.doubleclick.net/gampad/clk?id=241902991=/4140 ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
[Bacula-users] is 7.2 ready for prime time?
All, I typically patch bacula pretty frequently, but I saw the somewhat unusual notice on the latest release notes that warns it may not be ready for use in production. How stable is it? I don't really have the resources to test this out, but rather would have to go straight to production with it. I could always roll back, but that might entail the recovery from dump of a 900GB database. Opinions? thanks, Stephen -- Stephen Thompson Berkeley Seismological Laboratory step...@seismo.berkeley.edu215 McCone Hall # 4760 Office: 510.664.9177 University of California, Berkeley Remote: 510.214.6506 (Tue,Wed) Berkeley, CA 94720-4760 -- Monitor Your Dynamic Infrastructure at Any Scale With Datadog! Get real-time metrics from all of your servers, apps and tools in one place. SourceForge users - Click here to start your Free Trial of Datadog now! http://pubads.g.doubleclick.net/gampad/clk?id=241902991=/4140 ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] is 7.2 ready for prime time?
Help? Well, the compile and install went fine, but the update tables script is having issue. I was running 7.0.5 before. Not sure what database version, but likely whatever was appropriate to 7.0.5. First time I ran script as su'ed user which caused this... -- Altering mysql tables This script will update a Bacula MySQL database from version 12 to 14 which is needed to convert from Bacula Community version 5.0.x to 5.2.x ERROR 1045 (28000): Access denied for user 'stephen'@'localhost' (using password: YES) /home/bacula/conf/update_mysql_tables: line 31: [: !=: unary operator expected ERROR 1045 (28000): Access denied for user 'stephen'@'localhost' (using password: YES) Update of Bacula MySQL tables failed. -- I assumed because access denied, that the script failed entirely, but then running it again as proper user... Second time... --- ./update_bacula_tables Altering mysql tables This script will update a Bacula MySQL database from version 12 to 14 which is needed to convert from Bacula Community version 5.0.x to 5.2.x /home/bacula/conf/update_mysql_tables: line 31: [: too many arguments ERROR 1050 (42S01) at line 1: Table 'RestoreObject' already exists ERROR 1061 (42000) at line 17: Duplicate key name 'jobhisto_jobid_idx' ERROR 1060 (42S21) at line 19: Duplicate column name 'DeltaSeq' Update of Bacula MySQL tables succeeded. -- Seems like it either partially ran before or I had changes already present from 7.0.5 update. However, my Director will not start due to database version number not being 15, and if I run the script any more times... -- Altering mysql tables This script will update a Bacula MySQL database from version 12 to 14 which is needed to convert from Bacula Community version 5.0.x to 5.2.x The existing database is version 14 !! This script can only update an existing version 12 database to version 14. Error. Cannot upgrade this database. -- If it updatad the database to 14, why is it not able to update to 15 if that's what the Director requires? thanks! Stephen On 09/24/2015 11:21 AM, Kern Sibbald wrote: > Hello, > > We put a caution message in every release, particularly for new features > which are generally tested but not always tested in production. Normally > most of the issues turn up for non-Linux distributions where we either > have not tested or have tested less than Linux. > > Version 7.2.0 is as stable or more so than any prior major release. That > said, there are always a few minor problems for each release and this > one is no different. All the important problems (build issues on > Solaris and FreeBSD) have been corrected in the public git repository. > > Best regards, > Kern > > On 15-09-24 11:40 AM, Stephen Thompson wrote: >> All, >> >> I typically patch bacula pretty frequently, but I saw the somewhat >> unusual notice on the latest release notes that warns it may not be >> ready for use in production. How stable is it? I don't really have the >> resources to test this out, but rather would have to go straight to >> production with it. I could always roll back, but that might entail the >> recovery from dump of a 900GB database. Opinions? >> >> thanks, >> Stephen -- Stephen Thompson Berkeley Seismological Laboratory step...@seismo.berkeley.edu215 McCone Hall # 4760 Office: 510.664.9177 University of California, Berkeley Remote: 510.214.6506 (Tue,Wed) Berkeley, CA 94720-4760 -- ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] is 7.2 ready for prime time?
I run daily backups of my database and had finished my monthly full run for September, so I was technically covered. However I was not looking forward to restoring a 900+Gb mysql database from a text dump which on my system would take days, if not an entire week. The last time I had to restore database from backup it was 4 or so years ago and my database was only 300-400Gb back then. Stephen On 09/25/2015 08:50 AM, Raymond Burns Jr. wrote: > Did you run a backup of the database? > If not, I bet you were terrified with all the errors :) > Same thing happened to me going to 7.0.5, and it sent me for a frenzy. I > didn't run a backup of the database because of all the great responses > from people. > > When is the 7.2.0 rpm expected? Not running update until the rpm is there. > > On Fri, Sep 25, 2015 at 10:43 AM Stephen Thompson > <step...@seismo.berkeley.edu <mailto:step...@seismo.berkeley.edu>> wrote: > > > > Spoke too soon, I see what's going on, I was running update script from > new location (7.2.0) and it's referencing old location (7.0.5) and > running the wrong mysql script. > > > > On 09/25/2015 08:34 AM, Stephen Thompson wrote: > > > > Help? > > > > Well, the compile and install went fine, but the update tables > script is > > having issue. > > > > I was running 7.0.5 before. Not sure what database version, but > likely > > whatever was appropriate to 7.0.5. > > > > > > First time I ran script as su'ed user which caused this... > > -- > > Altering mysql tables > > > > This script will update a Bacula MySQL database from version 12 to 14 > >which is needed to convert from Bacula Community version 5.0.x > to 5.2.x > > > > ERROR 1045 (28000): Access denied for user 'stephen'@'localhost' > (using > > password: YES) > > /home/bacula/conf/update_mysql_tables: line 31: [: !=: unary operator > > expected > > ERROR 1045 (28000): Access denied for user 'stephen'@'localhost' > (using > > password: YES) > > Update of Bacula MySQL tables failed. > > -- > > > > I assumed because access denied, that the script failed entirely, but > > then running it again as proper user... > > > > Second time... > > --- > > ./update_bacula_tables > > Altering mysql tables > > > > This script will update a Bacula MySQL database from version 12 to 14 > >which is needed to convert from Bacula Community version 5.0.x > to 5.2.x > > > > /home/bacula/conf/update_mysql_tables: line 31: [: too many arguments > > ERROR 1050 (42S01) at line 1: Table 'RestoreObject' already exists > > ERROR 1061 (42000) at line 17: Duplicate key name > 'jobhisto_jobid_idx' > > ERROR 1060 (42S21) at line 19: Duplicate column name 'DeltaSeq' > > Update of Bacula MySQL tables succeeded. > > -- > > > > Seems like it either partially ran before or I had changes already > > present from 7.0.5 update. > > > > However, my Director will not start due to database version > number not > > being 15, and if I run the script any more times... > > -- > > Altering mysql tables > > > > This script will update a Bacula MySQL database from version 12 to 14 > >which is needed to convert from Bacula Community version 5.0.x > to 5.2.x > > > > > > The existing database is version 14 !! > > This script can only update an existing version 12 database to > version 14. > > Error. Cannot upgrade this database. > > -- > > > > > > If it updatad the database to 14, why is it not able to update to > 15 if > > that's what the Director requires? > > > > > > thanks! > > Stephen > > > > > > > > > > > > On 09/24/2015 11:21 AM, Kern Sibbald wrote: > >> Hello, > >> > >> We put a caution message in every release, particularly for new > features > >> which are generally tested but not always tested in production. > Normally > >> most of the issues turn up for non-Linux distributions where we > either > >> have not tested or have tested less than Linux. >
Re: [Bacula-users] is 7.2 ready for prime time?
Spoke too soon, I see what's going on, I was running update script from new location (7.2.0) and it's referencing old location (7.0.5) and running the wrong mysql script. On 09/25/2015 08:34 AM, Stephen Thompson wrote: > > Help? > > Well, the compile and install went fine, but the update tables script is > having issue. > > I was running 7.0.5 before. Not sure what database version, but likely > whatever was appropriate to 7.0.5. > > > First time I ran script as su'ed user which caused this... > -- > Altering mysql tables > > This script will update a Bacula MySQL database from version 12 to 14 >which is needed to convert from Bacula Community version 5.0.x to 5.2.x > > ERROR 1045 (28000): Access denied for user 'stephen'@'localhost' (using > password: YES) > /home/bacula/conf/update_mysql_tables: line 31: [: !=: unary operator > expected > ERROR 1045 (28000): Access denied for user 'stephen'@'localhost' (using > password: YES) > Update of Bacula MySQL tables failed. > -- > > I assumed because access denied, that the script failed entirely, but > then running it again as proper user... > > Second time... > --- > ./update_bacula_tables > Altering mysql tables > > This script will update a Bacula MySQL database from version 12 to 14 >which is needed to convert from Bacula Community version 5.0.x to 5.2.x > > /home/bacula/conf/update_mysql_tables: line 31: [: too many arguments > ERROR 1050 (42S01) at line 1: Table 'RestoreObject' already exists > ERROR 1061 (42000) at line 17: Duplicate key name 'jobhisto_jobid_idx' > ERROR 1060 (42S21) at line 19: Duplicate column name 'DeltaSeq' > Update of Bacula MySQL tables succeeded. > -- > > Seems like it either partially ran before or I had changes already > present from 7.0.5 update. > > However, my Director will not start due to database version number not > being 15, and if I run the script any more times... > -- > Altering mysql tables > > This script will update a Bacula MySQL database from version 12 to 14 >which is needed to convert from Bacula Community version 5.0.x to 5.2.x > > > The existing database is version 14 !! > This script can only update an existing version 12 database to version 14. > Error. Cannot upgrade this database. > -- > > > If it updatad the database to 14, why is it not able to update to 15 if > that's what the Director requires? > > > thanks! > Stephen > > > > > > On 09/24/2015 11:21 AM, Kern Sibbald wrote: >> Hello, >> >> We put a caution message in every release, particularly for new features >> which are generally tested but not always tested in production. Normally >> most of the issues turn up for non-Linux distributions where we either >> have not tested or have tested less than Linux. >> >> Version 7.2.0 is as stable or more so than any prior major release. That >> said, there are always a few minor problems for each release and this >> one is no different. All the important problems (build issues on >> Solaris and FreeBSD) have been corrected in the public git repository. >> >> Best regards, >> Kern >> >> On 15-09-24 11:40 AM, Stephen Thompson wrote: >>> All, >>> >>> I typically patch bacula pretty frequently, but I saw the somewhat >>> unusual notice on the latest release notes that warns it may not be >>> ready for use in production. How stable is it? I don't really have the >>> resources to test this out, but rather would have to go straight to >>> production with it. I could always roll back, but that might entail the >>> recovery from dump of a 900GB database. Opinions? >>> >>> thanks, >>> Stephen > -- Stephen Thompson Berkeley Seismological Laboratory step...@seismo.berkeley.edu215 McCone Hall # 4760 Office: 510.664.9177 University of California, Berkeley Remote: 510.214.6506 (Tue,Wed) Berkeley, CA 94720-4760 -- ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] is 7.2 ready for prime time?
So far so good. Minor snafu on my part when updating database, but I'm running 7.2 now. Looking good so far. Will find out more when hundreds of jobs run tonight. Stephen On 09/24/2015 08:40 AM, Stephen Thompson wrote: > > All, > > I typically patch bacula pretty frequently, but I saw the somewhat > unusual notice on the latest release notes that warns it may not be > ready for use in production. How stable is it? I don't really have the > resources to test this out, but rather would have to go straight to > production with it. I could always roll back, but that might entail the > recovery from dump of a 900GB database. Opinions? > > thanks, > Stephen > -- Stephen Thompson Berkeley Seismological Laboratory step...@seismo.berkeley.edu215 McCone Hall # 4760 Office: 510.664.9177 University of California, Berkeley Remote: 510.214.6506 (Tue,Wed) Berkeley, CA 94720-4760 -- ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] is 7.2 ready for prime time?
Thanks, I'll be upgrading soon. What known bugs are in the update_bacula_tables scripts? thanks, Stephen On 9/24/15 10:51 PM, Uwe Schuerkamp wrote: > On Thu, Sep 24, 2015 at 08:40:05AM -0700, Stephen Thompson wrote: >> >> All, >> >> I typically patch bacula pretty frequently, but I saw the somewhat >> unusual notice on the latest release notes that warns it may not be >> ready for use in production. How stable is it? I don't really have the >> resources to test this out, but rather would have to go straight to >> production with it. I could always roll back, but that might entail the >> recovery from dump of a 900GB database. Opinions? >> > > I upgraded five bacula instances of varying size over the last four > weeks or so, starting with the smallest (all were on 7.0.5 compiled > from source on CentOS), no issues so far apart from the little bugs in > the update_bacula_tables script. > > Cheers, Uwe > -- Stephen Thompson Berkeley Seismological Laboratory step...@seismo.berkeley.edu215 McCone Hall #4760 Office: 510.664.9177 University of California, Berkeley Remote: 510.214.6506 (Tue,Wed) Berkeley, CA 94720-4760 -- ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] 7.2 mysql issue?
Eric, Thanks for the reply. I've heard the postgres recommendation a fair number of times. A couple years back, we setup a parallel instance but even after tuning still wound up with _worse_ performance than with mysql. I could not figure out what to attribute this to (because it was in such contrast to all the pro-postgres recommendations) except possibly our memory-poor server - 8Gb RAM. At any rate, the only thing that's changed was the upgrade from 7.0.5 to 7.2.0. The table involved is definitely the File table. We do have jobs with 20-30 million records, so those jobs can be slow when it comes time for attribute insertion into the database (or to read out a file list for Accurate backups). This why we've historically had innodb lock timeout of 3600. However, it's only last week after the upgrade that we've ever had queries extend beyond that hour mark. We also went through a database cleaning process last month due to nearly reaching 1Tb and I can pretty authoritatively claim that we don't have orphan records. The database content and schema all appear to be appropriate. I was worried that queries had been rewritten that made it more efficient for other databases, but less so for mysql. More info... example from slow query logfile: # Time: 151001 1:28:14 # User@Host: bacula[bacula] @ localhost [] # Query_time: 3675.052083 Lock_time: 73.719795 Rows_sent: 0 Rows_examined: 3 SET timestamp=1443688094; INSERT INTO File (FileIndex, JobId, PathId, FilenameId, LStat, MD5, DeltaSeq) SELECT batch.FileIndex, batch.JobId, Path.PathId, Filename.FilenameId,batch.LStat, batch.MD5, batch.DeltaSeq FROM batch JOIN Path ON (batch.Path = Path.Path) JOIN Filename ON (batch.Name = Filename.Name); mysqld: mysql-5.1.73-5.el6_6.x86_64 record counts per table: File4,315,675,600 Filename154,748,787 Path28,534,411 innodb file sizes: 847708500 File.ibd 19488772Filename.ibd 8216580 Path.ibd 106500 PathHierarchy.ibd 57344 JobMedia.ibd 40960 PathVisibility.ibd 27648 Job.ibd 512 Media.ibd 176 FileSet.ibd 144 JobHisto.ibd 144 Client.ibd 112 RestoreObject.ibd 112 Pool.ibd 112 Log.ibd 112 BaseFiles.ibd 96 Version.ibd 96 UnsavedFiles.ibd 96 Storage.ibd 96 Status.ibd 96 MediaType.ibd 96 LocationLog.ibd 96 Location.ibd 96 Device.ibd 96 Counters.ibd 96 CDImages.ibd 4 Snapshot.MYI 0 Snapshot.MYD Not related, but I just noticed that somehow the new Snapshot table is MyISAM format. How did that happen? Regarding: > Would be nice also if you can give the number of Filename per Client (from the job table). Do you have a sample SQL to retrieve this stat? thanks, Stephen On 10/03/2015 12:02 AM, Eric Bollengier wrote: > Hello Stephen, > > On 10/03/2015 12:00 AM, Stephen Thompson wrote: >> >> >> All, >> >> I believe I'm having mysql database issues since upgrading to 7.2 (from >> 7.0.2). I run mysql innodb with 900Gb database that's largely the File >> table. > > For large catalog, we usually advise to use PostgreSQL where we have > multi-terabytes databases in production. > >> Since upgrading, I lose a few jobs a night due to database locking >> timeouts, which I have set to 3600. I also log slow queries. > > Can you get some information about these locks? On which table? Can you > give some statistics on your catalog like the size and the number of > records of the File, Filename and Path table? Would be nice also if you > can give the number of Filename per Client (from the job table). > > You might have many orphan Filenames, and MySQL is not always very good > to join large tables (it uses nested loops, and cannot use the index on > the Text column in all queries). > >> It appears that typically during a months I have about 90-100 queries >> that take longer than 15 minutes to run. Already this month (upgraded >> earlier this week), I have 32 queries that take longer than 15 minutes. >>At this rate (after 2 days) that will up my regular average of 90-100 >> to 480! >> >> Something is wrong and the coincidence is pretty strong that it's >> related to the upgrade. > > Maybe, but I'm not sure, we did not change a lot of thing in this area, > we did mostly refactoring. > > Best Regards, > Eric > -- Stephen Thompson Berkeley Seismological Laboratory step...@seismo.berkeley.edu215 McCone Hall # 4760 Office: 510.664.9177 University of California, Berkeley Remote: 510.214.6506 (Tue,Wed) Berkeley, CA 94720-4760 --
[Bacula-users] client initiated backups - bconsole vs tray?
All,
I've been trying to setup client initiated backups via FD remote=yes and
bconsole with no success. Regardless of the ACLs defined on Director,
the only command available on client's bconsole is "status" and even
that is the status of the local FD, not the DIR status. Every other
command yields...
2999 Invalid command
==MyConfigution==
server...
bacula-dir.conf:
Director {
Name = bacula-dir
DIRport = 9101
Password = "ABC123"
}
Console {
Name = dir-con-fd
Password = "DEF456"
CommandACL = *all*
ClientACL = *all*
JobACL = *all*
PoolACL = *all*
StorageACL = *all*
CatalogACL = *all*
FileSetACL = *all*
}
client...
bacula-fd.conf:
Director {
Name = bacula-dir
Password = "ABC123"
}
Console {
Name = dir-con-fd
DIRPort = 9101
Address =
Password = "DEF456"
}
Director {
Name = fd-con
Remote = yes
Password = "GHI789"
}
FileDaemon {
Name = bacula-fd
FDport = 9102
}
bconsole.conf:
Director {
Name = bacula-fd
DIRport = 9102
Address = localhost
Password = "NOT_USED-SEE_CONSOLE_SECTION"
}
Console {
Name = fd-con
Password = "GHI789"
}
I see the docs on this lean heavily toward tray. Does this even work
for bconsole? I saw a Kern comment that they use bconsole for testing
this feature, but I just cannot get it to let me run any command but a
local status of the FD.
Help?
thanks,
Stephen
--
Stephen Thompson Berkeley Seismo Lab
step...@seismo.berkeley.edu215 McCone Hall
Office: 510.664.9177 University of California
Remote: 510.214.6506 (Tue) Berkeley, CA 94720-4760
--
Check out the vibrant tech community on one of the world's most
engaging tech sites, Slashdot.org! http://sdm.link/slashdot
___
Bacula-users mailing list
Bacula-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] client initiated backups - bconsole vs tray?
Thanks Martin. That got me a step closer, but still not working. If I run bacula-fd in foreground, I can see that when I execute proxy command the FD outputs a successful connected to Director message. But running any other command under proxy in bconsole just hangs with no output from FD or from Director. Hmmm... Stephen On 7/5/18 8:21 AM, Martin Simmons wrote: On Tue, 3 Jul 2018 16:04:56 -0700, Stephen Thompson said: All, I've been trying to setup client initiated backups via FD remote=yes and bconsole with no success. Regardless of the ACLs defined on Director, the only command available on client's bconsole is "status" and even that is the status of the local FD, not the DIR status. Every other command yields... 2999 Invalid command You are not connected directly to the Director command loop after connecting bconsole to the local FD. According to the test (regress/tests/remote-console-test), you need to use the proxy command (without any arguments) to connect to the Director. __Martin -- Stephen Thompson Berkeley Seismo Lab step...@seismo.berkeley.edu215 McCone Hall Office: 510.664.9177 University of California Remote: 510.214.6506 (Tue) Berkeley, CA 94720-4760 -- Check out the vibrant tech community on one of the world's most engaging tech sites, Slashdot.org! http://sdm.link/slashdot ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] client initiated backups - bconsole vs tray?
I got this working without TLS enabled. Not sure why that breaks it, but perhaps something with how proxy is handled and how the TLS settings may not be applied properly for a remote=yes config, even though they are accepted as options. Oddly, even if I just enabled tls from the console to the FD, which allows for a local status of FD, this breaks proxy, even when the proxy connection does not have TLS enabled and works if the console to FD also has no TLS enabled. That the proxy connection might not be expecting TLS I can understand, but seems odd that a successful TLS connection from console to FD would break the FD's ability to proxy to the remote Director. Stephen On 7/6/18 2:44 PM, Stephen Thompson wrote: Well this led to unexpected results. Still 9.0.6, but running both FD and DIR in foreground with d900 both show startup messages, show console connecting to FD, show FD connecting to DIR when "proxy" is sent, but then when any command is sent and hangs, NEITHER FD NOR DIRECTOR output anything at all! 2000 OK Hello 214 Enter a period to cancel a command. *proxy 2000 proxy OK. *status FD output... fd: hello.c:262-0 Connecting to Director DIRECTOR:9101 fd: watchdog.c:197-0 Registered watchdog 7fc73401fa68, interval 15 one shot fd: btimers.c:145-0 Start thread timer 7fc73401d498 tid 7fc73bf4c700 for 15 secs. fd: bsock.c:237-0 Current A.B.C.D:9101 All W.X.Y.Z:9101 fd: bsock.c:166-0 who=Director daemon host=DIRECTOR port=9101 fd: bsock.c:349-0 OK connected to server Director daemon DIRECTOR:9101. fd: btimers.c:203-0 Stop thread timer 7fc73401d498 tid=7fc73bf4c700. fd: watchdog.c:217-0 Unregistered watchdog 7fc73401fa68 fd: watchdog.c:197-0 Registered watchdog 7fc73401d498, interval 15 one shot fd: btimers.c:177-0 Start bsock timer 7fc734005d18 tid=7fc73bf4c700 for 15 secs at 1530890871 fd: cram-md5.c:133-0 cram-get received: auth cram-md5 <195401314.1530890871@dir> ssl=2 fd: cram-md5.c:157-0 sending resp to challenge: jlJ1z7+S47xwcCkb2S+GGD fd: cram-md5.c:76-0 send: auth cram-md5 challenge <88308421.1530890871@fd> ssl=2 fd: cram-md5.c:95-0 Authenticate OK GD/TjH/8Dwc+4C0mJ8+2oD fd: tls.c:392-0 Check subject name name fd: bnet.c:280-0 TLS client negotiation established. fd: hello.c:335-0 >dird: 1000 OK auth fd: hello.c:342-0 November 2017) fd: hello.c:345-0 1000 OK: 103 DIRECTOR Version: 9.0.6 (20 November 2017) DIR output dir: bnet.c:569-0 socket=6 who=client host=A.B.C.D port=9101 dir: jcr.c:931-0 set_jcr_job_status(0, C) dir: jcr.c:940-0 OnEntry JobStatus=0 newJobstatus=C dir: jcr.c:951-0 Set new stat. old: 0,0 new: C,0 dir: jcr.c:956-0 leave setJobStatus old=0 new=C dir: job.c:1760-0 wstorage=STORAGE dir: job.c:1769-0 wstore=STORAGE where=Job resource dir: job.c:1429-0 JobId=0 created Job=-Console-.2018-07-06_08.27.51_05 dir: jcr.c:931-0 set_jcr_job_status(0, R) dir: jcr.c:940-0 OnEntry JobStatus=C newJobstatus=R dir: jcr.c:951-0 Set new stat. old: C,0 new: R,0 dir: jcr.c:956-0 leave setJobStatus old=C new=R dir: cram-md5.c:69-0 send: auth cram-md5 challenge <195401314.1530890871@dir> ssl=2 dir: cram-md5.c:133-0 cram-get received: auth cram-md5 <88308421.1530890871@fd> ssl=2 dir: cram-md5.c:157-0 sending resp to challenge: GD/TjH/8Dwc+4C0mJ8+2oD lawson-dir: bnet.c:230-0 TLS server negotiation established. I'm going to build 9.0.8 and see if I get different results. I believe I skipped TLS with the same results. Stephen On 7/6/18 7:23 AM, Stephen Thompson wrote: Yes, it does print 2000 proxy OK, but then in my case, the 'run' below would hang. And as I said, running the bacula-fd in the foreground shows a successful connection to Director when successful, but then nothing more. Also an unsuccessful connection (on purpose) is output form both the FD and the DIR, so they are definitely talking. Hmmm... I will try your foregrounded director suggestion. BTW, I'm also using TLS, which I'm hoping is not muddying the waters. Oh, and technically I'm running 9.0.6, so perhaps I should upgrade as well. Stephen On 7/6/18 3:50 AM, Martin Simmons wrote: It works for me in 9.0.8: Connecting to Director localhost:9102 2000 OK Hello 214 Enter a period to cancel a command. *proxy 2000 proxy OK. *run Automatically selected Catalog: MyCatalog Using Catalog "MyCatalog" A job name must be specified. The defined Job resources are: 1: Client1 ... Does it print "2000 proxy OK." and the "*" prompt after the proxy command? You could try running the Director in the foreground with -d900. __Martin On Thu, 5 Jul 2018 17:30:31 -0700, Stephen Thompson said: Thanks Martin. That got me a step closer, but still not working. If I run bacula-fd in foreground, I can see that when I execute proxy command the FD outputs a successful connected to Director message. But running any other command under proxy in bconsole just hangs with no output from FD or from Director. Hmmm... Ste
Re: [Bacula-users] possibly new mtx timing bug in 9x?
update... they may very well be hardware, though it did not seem like it at first. if it's a timing issue, it's not with the library but the drive. On 7/6/18 7:26 AM, Stephen Thompson wrote: Not sure if anyone else is seeing this, but sporadically, perhaps 2-3 times a month, after running various version of bacula on the same server/tape library for 10 years now and now running 9.0.6, we are seeing cases where bacula want's user intervention to mount a tape in a drive, but the tape is already in the drive AND bacula put it there. The only thing I can think is that the tape load step is somehow timing out and then not making the check to see whether the tape made it to the drive or not. thanks, Stephen -- Stephen Thompson Berkeley Seismo Lab step...@seismo.berkeley.edu215 McCone Hall Office: 510.664.9177 University of California Remote: 510.214.6506 (Tue) Berkeley, CA 94720-4760 -- Check out the vibrant tech community on one of the world's most engaging tech sites, Slashdot.org! http://sdm.link/slashdot ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] client initiated backups - bconsole vs tray?
Yes, it does print 2000 proxy OK, but then in my case, the 'run' below would hang. And as I said, running the bacula-fd in the foreground shows a successful connection to Director when successful, but then nothing more. Also an unsuccessful connection (on purpose) is output form both the FD and the DIR, so they are definitely talking. Hmmm... I will try your foregrounded director suggestion. BTW, I'm also using TLS, which I'm hoping is not muddying the waters. Oh, and technically I'm running 9.0.6, so perhaps I should upgrade as well. Stephen On 7/6/18 3:50 AM, Martin Simmons wrote: It works for me in 9.0.8: Connecting to Director localhost:9102 2000 OK Hello 214 Enter a period to cancel a command. *proxy 2000 proxy OK. *run Automatically selected Catalog: MyCatalog Using Catalog "MyCatalog" A job name must be specified. The defined Job resources are: 1: Client1 ... Does it print "2000 proxy OK." and the "*" prompt after the proxy command? You could try running the Director in the foreground with -d900. __Martin On Thu, 5 Jul 2018 17:30:31 -0700, Stephen Thompson said: Thanks Martin. That got me a step closer, but still not working. If I run bacula-fd in foreground, I can see that when I execute proxy command the FD outputs a successful connected to Director message. But running any other command under proxy in bconsole just hangs with no output from FD or from Director. Hmmm... Stephen On 7/5/18 8:21 AM, Martin Simmons wrote: On Tue, 3 Jul 2018 16:04:56 -0700, Stephen Thompson said: All, I've been trying to setup client initiated backups via FD remote=yes and bconsole with no success. Regardless of the ACLs defined on Director, the only command available on client's bconsole is "status" and even that is the status of the local FD, not the DIR status. Every other command yields... 2999 Invalid command You are not connected directly to the Director command loop after connecting bconsole to the local FD. According to the test (regress/tests/remote-console-test), you need to use the proxy command (without any arguments) to connect to the Director. __Martin -- Stephen Thompson Berkeley Seismo Lab step...@seismo.berkeley.edu215 McCone Hall Office: 510.664.9177 University of California Remote: 510.214.6506 (Tue) Berkeley, CA 94720-4760 -- Stephen Thompson Berkeley Seismo Lab step...@seismo.berkeley.edu215 McCone Hall Office: 510.664.9177 University of California Remote: 510.214.6506 (Tue) Berkeley, CA 94720-4760 -- Check out the vibrant tech community on one of the world's most engaging tech sites, Slashdot.org! http://sdm.link/slashdot ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
[Bacula-users] possibly new mtx timing bug in 9x?
Not sure if anyone else is seeing this, but sporadically, perhaps 2-3 times a month, after running various version of bacula on the same server/tape library for 10 years now and now running 9.0.6, we are seeing cases where bacula want's user intervention to mount a tape in a drive, but the tape is already in the drive AND bacula put it there. The only thing I can think is that the tape load step is somehow timing out and then not making the check to see whether the tape made it to the drive or not. thanks, Stephen -- Stephen Thompson Berkeley Seismo Lab step...@seismo.berkeley.edu215 McCone Hall Office: 510.664.9177 University of California Remote: 510.214.6506 (Tue) Berkeley, CA 94720-4760 -- Check out the vibrant tech community on one of the world's most engaging tech sites, Slashdot.org! http://sdm.link/slashdot ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] client initiated backups - bconsole vs tray?
Well this led to unexpected results. Still 9.0.6, but running both FD and DIR in foreground with d900 both show startup messages, show console connecting to FD, show FD connecting to DIR when "proxy" is sent, but then when any command is sent and hangs, NEITHER FD NOR DIRECTOR output anything at all! 2000 OK Hello 214 Enter a period to cancel a command. *proxy 2000 proxy OK. *status FD output... fd: hello.c:262-0 Connecting to Director DIRECTOR:9101 fd: watchdog.c:197-0 Registered watchdog 7fc73401fa68, interval 15 one shot fd: btimers.c:145-0 Start thread timer 7fc73401d498 tid 7fc73bf4c700 for 15 secs. fd: bsock.c:237-0 Current A.B.C.D:9101 All W.X.Y.Z:9101 fd: bsock.c:166-0 who=Director daemon host=DIRECTOR port=9101 fd: bsock.c:349-0 OK connected to server Director daemon DIRECTOR:9101. fd: btimers.c:203-0 Stop thread timer 7fc73401d498 tid=7fc73bf4c700. fd: watchdog.c:217-0 Unregistered watchdog 7fc73401fa68 fd: watchdog.c:197-0 Registered watchdog 7fc73401d498, interval 15 one shot fd: btimers.c:177-0 Start bsock timer 7fc734005d18 tid=7fc73bf4c700 for 15 secs at 1530890871 fd: cram-md5.c:133-0 cram-get received: auth cram-md5 <195401314.1530890871@dir> ssl=2 fd: cram-md5.c:157-0 sending resp to challenge: jlJ1z7+S47xwcCkb2S+GGD fd: cram-md5.c:76-0 send: auth cram-md5 challenge <88308421.1530890871@fd> ssl=2 fd: cram-md5.c:95-0 Authenticate OK GD/TjH/8Dwc+4C0mJ8+2oD fd: tls.c:392-0 Check subject name name fd: bnet.c:280-0 TLS client negotiation established. fd: hello.c:335-0 >dird: 1000 OK auth fd: hello.c:342-0 November 2017) fd: hello.c:345-0 1000 OK: 103 DIRECTOR Version: 9.0.6 (20 November 2017) DIR output dir: bnet.c:569-0 socket=6 who=client host=A.B.C.D port=9101 dir: jcr.c:931-0 set_jcr_job_status(0, C) dir: jcr.c:940-0 OnEntry JobStatus=0 newJobstatus=C dir: jcr.c:951-0 Set new stat. old: 0,0 new: C,0 dir: jcr.c:956-0 leave setJobStatus old=0 new=C dir: job.c:1760-0 wstorage=STORAGE dir: job.c:1769-0 wstore=STORAGE where=Job resource dir: job.c:1429-0 JobId=0 created Job=-Console-.2018-07-06_08.27.51_05 dir: jcr.c:931-0 set_jcr_job_status(0, R) dir: jcr.c:940-0 OnEntry JobStatus=C newJobstatus=R dir: jcr.c:951-0 Set new stat. old: C,0 new: R,0 dir: jcr.c:956-0 leave setJobStatus old=C new=R dir: cram-md5.c:69-0 send: auth cram-md5 challenge <195401314.1530890871@dir> ssl=2 dir: cram-md5.c:133-0 cram-get received: auth cram-md5 <88308421.1530890871@fd> ssl=2 dir: cram-md5.c:157-0 sending resp to challenge: GD/TjH/8Dwc+4C0mJ8+2oD lawson-dir: bnet.c:230-0 TLS server negotiation established. I'm going to build 9.0.8 and see if I get different results. I believe I skipped TLS with the same results. Stephen On 7/6/18 7:23 AM, Stephen Thompson wrote: Yes, it does print 2000 proxy OK, but then in my case, the 'run' below would hang. And as I said, running the bacula-fd in the foreground shows a successful connection to Director when successful, but then nothing more. Also an unsuccessful connection (on purpose) is output form both the FD and the DIR, so they are definitely talking. Hmmm... I will try your foregrounded director suggestion. BTW, I'm also using TLS, which I'm hoping is not muddying the waters. Oh, and technically I'm running 9.0.6, so perhaps I should upgrade as well. Stephen On 7/6/18 3:50 AM, Martin Simmons wrote: It works for me in 9.0.8: Connecting to Director localhost:9102 2000 OK Hello 214 Enter a period to cancel a command. *proxy 2000 proxy OK. *run Automatically selected Catalog: MyCatalog Using Catalog "MyCatalog" A job name must be specified. The defined Job resources are: 1: Client1 ... Does it print "2000 proxy OK." and the "*" prompt after the proxy command? You could try running the Director in the foreground with -d900. __Martin On Thu, 5 Jul 2018 17:30:31 -0700, Stephen Thompson said: Thanks Martin. That got me a step closer, but still not working. If I run bacula-fd in foreground, I can see that when I execute proxy command the FD outputs a successful connected to Director message. But running any other command under proxy in bconsole just hangs with no output from FD or from Director. Hmmm... Stephen On 7/5/18 8:21 AM, Martin Simmons wrote: On Tue, 3 Jul 2018 16:04:56 -0700, Stephen Thompson said: All, I've been trying to setup client initiated backups via FD remote=yes and bconsole with no success. Regardless of the ACLs defined on Director, the only command available on client's bconsole is "status" and even that is the status of the local FD, not the DIR status. Every other command yields... 2999 Invalid command You are not connected directly to the Director command loop after connecting bconsole to the local FD. According to the test (regress/tests/remote-console-test), you need to use the proxy command (without any arguments) to connect to the Director. __Martin -- Stephen Thompson
[Bacula-users] can file retention be job rather than client based?
I believe the answer is no, but as a happy bacula user for 10 years I am somewhat surprised at the lack of flexibility. The scenarios is this: A fileserver (1 client) with dozens of large (size-wise) filesystems (12 jobs), but a couple of those filesystems are large (filecount-wise). We would really like to set different file retention periods on those high-filecount jobs (50+million), because they are forcing the Catalog to go beyond our size constraints. However, we also don't want to lose the file retention longevity of that client's other jobs (5 years). The only hack I can think of is to define 2 clients for 1 actual host, but I'd rather not go down that route, because tracking jobs and associating them, especially over multiple years, will get that much more tricky. Ideas? thanks, Stephen -- Stephen Thompson Berkeley Seismo Lab step...@seismo.berkeley.edu215 McCone Hall Office: 510.664.9177 University of California Berkeley, CA 94720-4760 -- Check out the vibrant tech community on one of the world's most engaging tech sites, Slashdot.org! http://sdm.link/slashdot ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] can file retention be job rather than client based?
Thanks Kern. I think given the limited nature of his need, I may use a postrun script to simply wipe database records out of band. Also if I did use multi-client definitions, I would need to use the same pool as they all go to the same monthly tapes. Stephen On 4/10/18 11:59 PM, Kern Sibbald wrote: Hello Stephen, What you are asking for, as you suspect, does not exist and implementing it would be a bit problematic because every Job would need to keep it's own retention period. For one client, there can be any number of Jobs -- typically thousands. Thus the catalog would grow faster (more data for the File table having the most records), and the complexity of pruning including the time to prune would probably explode -- probably thousands of times slower. I have never used two Client definitions to backup the same machine, but in principle it would work fine. If you name your Clients appropriately it might be easier to remember what was done. E.g. Client1-Normal-Files, Client1-Archived-Files, ... Also, if you put clear comments on the resource definitions, it would help. Note two things, if you go this route: 1. Be sure to define each of your two Client1-xxx with different Pools with different Volume retention periods 2. I would appreciate feedback on how this works -- especially operationally Best regards, Kern PS: At the current time the Enterprise version of Bacula has a number of performance improvements that should significantly speed up the backups of 50+million files. It does this at a small extra expense (size) of the catalog. On 04/07/2018 06:21 AM, Stephen Thompson wrote: I believe the answer is no, but as a happy bacula user for 10 years I am somewhat surprised at the lack of flexibility. The scenarios is this: A fileserver (1 client) with dozens of large (size-wise) filesystems (12 jobs), but a couple of those filesystems are large (filecount-wise). We would really like to set different file retention periods on those high-filecount jobs (50+million), because they are forcing the Catalog to go beyond our size constraints. However, we also don't want to lose the file retention longevity of that client's other jobs (5 years). The only hack I can think of is to define 2 clients for 1 actual host, but I'd rather not go down that route, because tracking jobs and associating them, especially over multiple years, will get that much more tricky. Ideas? thanks, Stephen -- Stephen Thompson Berkeley Seismo Lab step...@seismo.berkeley.edu215 McCone Hall Office: 510.664.9177 University of California Remote: 510.214.6506 (Tue) Berkeley, CA 94720-4760 -- Check out the vibrant tech community on one of the world's most engaging tech sites, Slashdot.org! http://sdm.link/slashdot ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] can file retention be job rather than client based?
Thanks... Yeah, I'm leaning towards a post or pre job script that actually prunes (or more likely purges) the file records I need to jettison. Stephen On 4/7/18 3:38 AM, Heitor Faria wrote: Hello Stefaphen, I believe the answer is no, but as a happy bacula user for 10 years I am somewhat surprised at the lack of flexibility. Alternative solutions with proprietary catalog data are much more inflexible. The scenarios is this: A fileserver (1 client) with dozens of large (size-wise) filesystems (12 jobs), but a couple of those filesystems are large (filecount-wise). We would really like to set different file retention periods on those high-filecount jobs (50+million), because they are forcing the Catalog to go beyond our size constraints. However, we also don't want to lose the file retention longevity of that client's other jobs (5 years). The only hack I can think of is to define 2 clients for 1 actual host, but I'd rather not go down that route, because tracking jobs and associating them, especially over multiple years, will get that much more tricky. File & Job Retention can be set in Pool resource instead of Client one. You can also try to modify the prior used manual Bacula pruning Perl script to only prune files you need and not delete the whole job <http://blog.bacula.org/whitepapers/manual_prune.pl>. Finally, you can even use dynamically generated Filesets with lower or greater file sizes to automate their backup distribution. <http://bacula.us/bacula-fileset-on-client-configuration-remote-fileset/> Ideas? thanks, Stephen -- Regards, -- Stephen Thompson Berkeley Seismo Lab step...@seismo.berkeley.edu215 McCone Hall Office: 510.664.9177 University of California Remote: 510.214.6506 (Tue) Berkeley, CA 94720-4760 -- Check out the vibrant tech community on one of the world's most engaging tech sites, Slashdot.org! http://sdm.link/slashdot ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
[Bacula-users] which database version for bacula 9.0.6?
I'm a little confused. The release notes for 9.0.0 (and possibly all 9x) say that a database upgrade is require, and to run the update_bacula_table script. I am running 7.4.4 at the moment and my database version (from version table) says I'm at 15, but the script to update mysql tables with bacula 9.0.6 apparently upgrades database to 15. Does this sound right? How can I already be a version that's for 9x, or did 15 come during 7x, which is why I have it, and the note in 9x release notes about needing database upgrade is because it can't hurt to run the script and many non-9x installations might not yet be at 15? Is the database version and how it corresponds to bacula versions documented in a list anywhere? I ask, because I want to know ahead of time how risky this is. I do backup my catalog, but it could take a week to restore from backup, so I just want to know whether I really will doing a database upgrade or not. thanks! Stephen -- Stephen Thompson Berkeley Seismo Lab step...@seismo.berkeley.edu215 McCone Hall Office: 510.664.9177 University of California Remote: 510.214.6506 (Tue) Berkeley, CA 94720-4760 -- Check out the vibrant tech community on one of the world's most engaging tech sites, Slashdot.org! http://sdm.link/slashdot ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
[Bacula-users] PurgedFiles in Job tables- Not toggled when Volumes are purged?
In looking at doing this out of band (not pruning feature) I've run into a tracking snag. We tend to purge volumes manually after a year when we want to reuse them, but it looks like purging volumes does not change the PurgedFiles column in the Job table for the Job's that have had their Files Purged. It appears that only happens if the Files are purged at the Job level. Can anyone confirm that is expected behaviour? I may need to purge all the Jobs on a volume, before I purge the volume, in order to get the flags set properly, so that I can more easily track which Job's have had their Files purged and which have not. Stephen On 04/11/2018 06:25 AM, Stephen Thompson wrote: Thanks Kern. I think given the limited nature of his need, I may use a postrun script to simply wipe database records out of band. Also if I did use multi-client definitions, I would need to use the same pool as they all go to the same monthly tapes. Stephen On 4/10/18 11:59 PM, Kern Sibbald wrote: Hello Stephen, What you are asking for, as you suspect, does not exist and implementing it would be a bit problematic because every Job would need to keep it's own retention period. For one client, there can be any number of Jobs -- typically thousands. Thus the catalog would grow faster (more data for the File table having the most records), and the complexity of pruning including the time to prune would probably explode -- probably thousands of times slower. I have never used two Client definitions to backup the same machine, but in principle it would work fine. If you name your Clients appropriately it might be easier to remember what was done. E.g. Client1-Normal-Files, Client1-Archived-Files, ... Also, if you put clear comments on the resource definitions, it would help. Note two things, if you go this route: 1. Be sure to define each of your two Client1-xxx with different Pools with different Volume retention periods 2. I would appreciate feedback on how this works -- especially operationally Best regards, Kern PS: At the current time the Enterprise version of Bacula has a number of performance improvements that should significantly speed up the backups of 50+million files. It does this at a small extra expense (size) of the catalog. On 04/07/2018 06:21 AM, Stephen Thompson wrote: I believe the answer is no, but as a happy bacula user for 10 years I am somewhat surprised at the lack of flexibility. The scenarios is this: A fileserver (1 client) with dozens of large (size-wise) filesystems (12 jobs), but a couple of those filesystems are large (filecount-wise). We would really like to set different file retention periods on those high-filecount jobs (50+million), because they are forcing the Catalog to go beyond our size constraints. However, we also don't want to lose the file retention longevity of that client's other jobs (5 years). The only hack I can think of is to define 2 clients for 1 actual host, but I'd rather not go down that route, because tracking jobs and associating them, especially over multiple years, will get that much more tricky. Ideas? thanks, Stephen -- Stephen Thompson Berkeley Seismo Lab step...@seismo.berkeley.edu215 McCone Hall Office: 510.664.9177 University of California Remote: 510.214.6506 (Tue) Berkeley, CA 94720-4760 -- Check out the vibrant tech community on one of the world's most engaging tech sites, Slashdot.org! http://sdm.link/slashdot ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] which database version for bacula 9.0.6?
Nevermind. I was looking in wrong place. I see that 9x involves an update from 15 to 16 and the script to do that with. Sorry, though I still wonder if there's a mapping somewhere that lists db versions against bacula versions. Stephen On 4/14/18 7:44 PM, Stephen Thompson wrote: I'm a little confused. The release notes for 9.0.0 (and possibly all 9x) say that a database upgrade is require, and to run the update_bacula_table script. I am running 7.4.4 at the moment and my database version (from version table) says I'm at 15, but the script to update mysql tables with bacula 9.0.6 apparently upgrades database to 15. Does this sound right? How can I already be a version that's for 9x, or did 15 come during 7x, which is why I have it, and the note in 9x release notes about needing database upgrade is because it can't hurt to run the script and many non-9x installations might not yet be at 15? Is the database version and how it corresponds to bacula versions documented in a list anywhere? I ask, because I want to know ahead of time how risky this is. I do backup my catalog, but it could take a week to restore from backup, so I just want to know whether I really will doing a database upgrade or not. thanks! Stephen -- Stephen Thompson Berkeley Seismo Lab step...@seismo.berkeley.edu215 McCone Hall Office: 510.664.9177 University of California Remote: 510.214.6506 (Tue) Berkeley, CA 94720-4760 -- Check out the vibrant tech community on one of the world's most engaging tech sites, Slashdot.org! http://sdm.link/slashdot ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] PurgedFiles in Job tables- Not toggled when Volumes are purged?
Looks like I may have been seeing things like Canceled (A) jobs that never had any File records to begin with, and therefore were never deleted with Volume purging and still have PurgedFiles still set to 0. Stephen On 04/19/2018 09:53 AM, Martin Simmons wrote: The "purge volumes" command deletes the job records, so there is no row anymore in which to set PurgedFiles. What is the exact bconsole command line you are running to purge volumes? __Martin On Mon, 16 Apr 2018 09:25:11 -0700, Stephen Thompson said: In looking at doing this out of band (not pruning feature) I've run into a tracking snag. We tend to purge volumes manually after a year when we want to reuse them, but it looks like purging volumes does not change the PurgedFiles column in the Job table for the Job's that have had their Files Purged. It appears that only happens if the Files are purged at the Job level. Can anyone confirm that is expected behaviour? I may need to purge all the Jobs on a volume, before I purge the volume, in order to get the flags set properly, so that I can more easily track which Job's have had their Files purged and which have not. Stephen On 04/11/2018 06:25 AM, Stephen Thompson wrote: Thanks Kern. I think given the limited nature of his need, I may use a postrun script to simply wipe database records out of band. Also if I did use multi-client definitions, I would need to use the same pool as they all go to the same monthly tapes. Stephen On 4/10/18 11:59 PM, Kern Sibbald wrote: Hello Stephen, What you are asking for, as you suspect, does not exist and implementing it would be a bit problematic because every Job would need to keep it's own retention period. For one client, there can be any number of Jobs -- typically thousands. Thus the catalog would grow faster (more data for the File table having the most records), and the complexity of pruning including the time to prune would probably explode -- probably thousands of times slower. I have never used two Client definitions to backup the same machine, but in principle it would work fine. If you name your Clients appropriately it might be easier to remember what was done. E.g. Client1-Normal-Files, Client1-Archived-Files, ... Also, if you put clear comments on the resource definitions, it would help. Note two things, if you go this route: 1. Be sure to define each of your two Client1-xxx with different Pools with different Volume retention periods 2. I would appreciate feedback on how this works -- especially operationally Best regards, Kern PS: At the current time the Enterprise version of Bacula has a number of performance improvements that should significantly speed up the backups of 50+million files. It does this at a small extra expense (size) of the catalog. On 04/07/2018 06:21 AM, Stephen Thompson wrote: I believe the answer is no, but as a happy bacula user for 10 years I am somewhat surprised at the lack of flexibility. The scenarios is this: A fileserver (1 client) with dozens of large (size-wise) filesystems (12 jobs), but a couple of those filesystems are large (filecount-wise). We would really like to set different file retention periods on those high-filecount jobs (50+million), because they are forcing the Catalog to go beyond our size constraints. However, we also don't want to lose the file retention longevity of that client's other jobs (5 years). The only hack I can think of is to define 2 clients for 1 actual host, but I'd rather not go down that route, because tracking jobs and associating them, especially over multiple years, will get that much more tricky. Ideas? thanks, Stephen -- Stephen Thompson Berkeley Seismo Lab step...@seismo.berkeley.edu215 McCone Hall Office: 510.664.9177 University of California Remote: 510.214.6506 (Tue) Berkeley, CA 94720-4760 -- Check out the vibrant tech community on one of the world's most engaging tech sites, Slashdot.org! http://sdm.link/slashdot ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users -- Stephen Thompson Berkeley Seismo Lab step...@seismo.berkeley.edu215 McCone Hall Office: 510.664.9177 University of California Remote: 510.214.6506 (Tue) Berkeley, CA 94720-4760 -- Check out the vibrant tech community on one of the world's most engaging tech sites, Slashdot.org! http://sdm.link/slashdot ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] Packet size too big (NOT a version mismatch)
However, even just backing up /Users results in... 04-Jan 11:31 SD JobId 88: Fatal error: bsock.c:530 Packet size=1387166 too big from "client:1.2.3.4:9103". Maximum permitted 100. Terminating connection. Stephen On 1/4/22 11:26 AM, Stephen Thompson wrote: Yes, backing up a single file on my problem hosts does succeed. H... Stephen On 1/4/22 11:23 AM, Stephen Thompson wrote: That's a good test, which I apparently have not tried. I will do so. thanks, Stephen On 1/4/22 11:20 AM, Martin Simmons wrote: Is this happening for all backups? What happens if you run a backup with a minimal fileset that lists just one small file? __Martin On Tue, 4 Jan 2022 08:13:46 -0800, Stephen Thompson said: I am still seeing the same issue on Monterey as on Big Sur with 11.0.5 compiled from source and CoreFoundation linked in. 04-Jan 07:56 SD JobId 88: Fatal error: bsock.c:530 Packet size=1387165 too big from "client:1.2.3.4:9103". Maximum permitted 100. Terminating connection. Stephen On Tue, Jan 4, 2022 at 7:02 AM Stephen Thompson < stephen.thomp...@berkeley.edu> wrote: Graham, Thanks for presenting Monterey as a possibility! I am seeing the same issue under Monterrey as I have under Big Sur, but to know someone else does not means that it's possible. I should double check that I am using a freshly compiled client on Monterey and not just the one that I compiled on Big Sur. I am backing up Macs with bacula, but not really for system recovery, more to backup user files/documents that they may not be backing up themselves. I do note a number of Mac system files that refuse to be backed up, but again for my purposes, I do not care too much. It would be nice to be able to BMR a Mac, but not a requirement where I am at, being operationally a Linux shop. Stephen On Tue, Jan 4, 2022 at 6:20 AM Graham Sparks wrote: Hi David, I use Time Machine (for the System disk) as well as Bacula on my Mac, as I'd still need the Time Machine backup to do a bare-metal restore (with Apps). I use Bacula to back up this and an external data drive. Rather than purchasing a separate "Time Capsule", I set up Samba on a Linux VM to expose an SMB share that the Mac sees as a Time Capsule drive ( https://wiki.samba.org/index.php/Configure_Samba_to_Work_Better_with_Mac_OS_X ). I had one problem with Time Machine a few months ago, where it stopped backing up data and insisted on starting the backup 'chain' from scratch again. I was a little miffed . I'm afraid I can only confirm that the Bacula v9.6 and v11 file daemons worked for me under macOS Catalina and Monetery (I skipped Big Sur. Not for good reason---just laziness). Both v9 and v11 clients were compiled from source (setting the linker flags to "-framework CoreFoundation" as already suggested). I've personally not run in to problems with System Integrity Protection, although I do give the bacula-fd executable "Full Disk" permissions. Thanks. -- Graham Sparks From: David Brodbeck Sent: 03 January 2022 18:36 Cc: bacula-users@lists.sourceforge.net < bacula-users@lists.sourceforge.net> Subject: Re: [Bacula-users] Packet size too big (NOT a version mismatch) I'm curious if anyone has moved away from Bacula on macOS and what alternatives they're using. Even before this, it was getting more and more awkward to set up -- bacula really doesn't play well with SIP, for example, and running "csrutil disable" on every system is not a security best practice. On Wed, Dec 8, 2021 at 4:46 PM Stephen Thompson < stephen.thomp...@berkeley.edu> wrote: Disappointing... I am having the same issue on BigSur with the 11.0.5 release as I had with 9x. 08-Dec 15:42 SD JobId 878266: Fatal error: bsock.c:530 Packet size=1387166 too big from "client:1.2.3.4:8103". Maximum permitted 100. Terminating connection. Setting 'Maximum Network Buffer Size' does not appear to solve issue. Are there users out there successfully running a bacula client on Big Sur?? Stephen On Wed, Dec 1, 2021 at 3:25 PM Stephen Thompson < stephen.thomp...@berkeley.edu> wrote: Not sure if this is correct, but I've been able to at least compile bacula client 11.0.5 on Big Sur by doing before configure step: LDFLAGS='-framework CoreFoundation' We'll see next up whether it runs and whether it exhibits the issue seen under Big Sur for 9x client. Stephen On Tue, Nov 23, 2021 at 7:32 AM Stephen Thompson < stephen.thomp...@berkeley.edu> wrote: Josh, Thanks for the tip. That did not appear to be the cause of this issue, though perhaps it will fix a yet to be found issue that I would have run into after I get past this compilation error. Stephen On Mon, Nov 22, 2021 at 9:22 AM Josh Fisher wrote: On 11/22/21 10:46, Stephen Thompson wrote: All, I too was having the issue with running a 9x client on Big
Re: [Bacula-users] Packet size too big (NOT a version mismatch)
Thanks. I have large file support off, though I am not sure that's intentional. I will double check that. On 1/4/22 11:55 AM, Graham Sparks wrote: I'm afraid I don't enable encryption in my backup jobs (I know I should ) so I don't know if that causes an issue. I'll have a quick look some time to see what happens when I enable encryption. I think I've reached my limit here, but it might be worth checking the following file to make sure all the compilation options took successfully (thinking aloud here, but "Large File Support" caught my attention): $BHOME/bin/bacula_config Thanks. -- Stephen Thompson Berkeley Seismology Lab stephen.thomp...@berkeley.edu 307 McCone Hall Office: 510.664.9177 University of California Remote: 510.214.6506 (Tue) Berkeley, CA 94720-4760 ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] Packet size too big (NOT a version mismatch)
Yes, backing up a single file on my problem hosts does succeed. H... Stephen On 1/4/22 11:23 AM, Stephen Thompson wrote: That's a good test, which I apparently have not tried. I will do so. thanks, Stephen On 1/4/22 11:20 AM, Martin Simmons wrote: Is this happening for all backups? What happens if you run a backup with a minimal fileset that lists just one small file? __Martin On Tue, 4 Jan 2022 08:13:46 -0800, Stephen Thompson said: I am still seeing the same issue on Monterey as on Big Sur with 11.0.5 compiled from source and CoreFoundation linked in. 04-Jan 07:56 SD JobId 88: Fatal error: bsock.c:530 Packet size=1387165 too big from "client:1.2.3.4:9103". Maximum permitted 100. Terminating connection. Stephen On Tue, Jan 4, 2022 at 7:02 AM Stephen Thompson < stephen.thomp...@berkeley.edu> wrote: Graham, Thanks for presenting Monterey as a possibility! I am seeing the same issue under Monterrey as I have under Big Sur, but to know someone else does not means that it's possible. I should double check that I am using a freshly compiled client on Monterey and not just the one that I compiled on Big Sur. I am backing up Macs with bacula, but not really for system recovery, more to backup user files/documents that they may not be backing up themselves. I do note a number of Mac system files that refuse to be backed up, but again for my purposes, I do not care too much. It would be nice to be able to BMR a Mac, but not a requirement where I am at, being operationally a Linux shop. Stephen On Tue, Jan 4, 2022 at 6:20 AM Graham Sparks wrote: Hi David, I use Time Machine (for the System disk) as well as Bacula on my Mac, as I'd still need the Time Machine backup to do a bare-metal restore (with Apps). I use Bacula to back up this and an external data drive. Rather than purchasing a separate "Time Capsule", I set up Samba on a Linux VM to expose an SMB share that the Mac sees as a Time Capsule drive ( https://wiki.samba.org/index.php/Configure_Samba_to_Work_Better_with_Mac_OS_X ). I had one problem with Time Machine a few months ago, where it stopped backing up data and insisted on starting the backup 'chain' from scratch again. I was a little miffed . I'm afraid I can only confirm that the Bacula v9.6 and v11 file daemons worked for me under macOS Catalina and Monetery (I skipped Big Sur. Not for good reason---just laziness). Both v9 and v11 clients were compiled from source (setting the linker flags to "-framework CoreFoundation" as already suggested). I've personally not run in to problems with System Integrity Protection, although I do give the bacula-fd executable "Full Disk" permissions. Thanks. -- Graham Sparks From: David Brodbeck Sent: 03 January 2022 18:36 Cc: bacula-users@lists.sourceforge.net < bacula-users@lists.sourceforge.net> Subject: Re: [Bacula-users] Packet size too big (NOT a version mismatch) I'm curious if anyone has moved away from Bacula on macOS and what alternatives they're using. Even before this, it was getting more and more awkward to set up -- bacula really doesn't play well with SIP, for example, and running "csrutil disable" on every system is not a security best practice. On Wed, Dec 8, 2021 at 4:46 PM Stephen Thompson < stephen.thomp...@berkeley.edu> wrote: Disappointing... I am having the same issue on BigSur with the 11.0.5 release as I had with 9x. 08-Dec 15:42 SD JobId 878266: Fatal error: bsock.c:530 Packet size=1387166 too big from "client:1.2.3.4:8103". Maximum permitted 100. Terminating connection. Setting 'Maximum Network Buffer Size' does not appear to solve issue. Are there users out there successfully running a bacula client on Big Sur?? Stephen On Wed, Dec 1, 2021 at 3:25 PM Stephen Thompson < stephen.thomp...@berkeley.edu> wrote: Not sure if this is correct, but I've been able to at least compile bacula client 11.0.5 on Big Sur by doing before configure step: LDFLAGS='-framework CoreFoundation' We'll see next up whether it runs and whether it exhibits the issue seen under Big Sur for 9x client. Stephen On Tue, Nov 23, 2021 at 7:32 AM Stephen Thompson < stephen.thomp...@berkeley.edu> wrote: Josh, Thanks for the tip. That did not appear to be the cause of this issue, though perhaps it will fix a yet to be found issue that I would have run into after I get past this compilation error. Stephen On Mon, Nov 22, 2021 at 9:22 AM Josh Fisher wrote: On 11/22/21 10:46, Stephen Thompson wrote: All, I too was having the issue with running a 9x client on Big Sur. I've tried compiling 11.0.5 but have not found my way past: This might be due to a libtool.m4 bug having to do with MacOS changing the major Darwin version from 19.x to 20.x. There is a patch at https://www.mail-archive.com/libtool-patches@gnu.org/msg07396.html Linking bacula-fd ... /U
Re: [Bacula-users] Packet size too big (NOT a version mismatch)
Graham, Thanks. I am confident that it's not a networking issue (at least one external to the Macs). The new problem only shows on hosts that have been updated to Big Sur or Monterey (with or without rebuilt client, both 9x and 11s). High Sierra and earlier hosts never yield the 'too big... Maximum permitted 100' error, but Big Sur/Monterey always do. I use xcode along with homebrew openssl 1.1. To further describe, the BigSur/Monterrey host jobs do partially complete, successfully sending many GBs of data and files, including a few warnings about unreadable system files, but ultimately the jobs crap out with the same error. My build options... BHOME=/Users/bacula EMAIL=bacula@DOMAINNAME env CFLAGS='-g -O2' LDFLAGS='-framework CoreFoundation' \ ./configure \ --prefix=$BHOME \ --sbindir=$BHOME/bin \ --sysconfdir=$BHOME/conf \ --with-working-dir=$BHOME/work \ --with-archivedir=$BHOME/archive \ --with-bsrdir=$BHOME/log \ --with-logdir=$BHOME/log \ --with-pid-dir=/var/run \ --with-subsys-dir=/var/run \ --with-basename=SERVER \ --with-hostname=SERVER.DOMAINNAME \ --with-dump-email=$EMAIL \ --with-openssl=/usr/local/opt/openssl\@1.1 \ --enable-smartalloc \ --disable-readline \ --enable-conio \ --enable-client-only \ | tee configure.out thanks again, Stephen On 1/4/22 10:54 AM, Graham Sparks wrote: Hi Stephen, I've had a quick read of the archive (I'm late to the mailing list party) and see you've tried lots, so I'll try to say something constructive. I tried to recreate the packet size error, crudely, by directing the Bacula server to a web page instead of the client FD (incidentally, this recreates it well). Therefore, I think it's worth making sure the server and client are communicating without interruption, just in case something else is being returned (perhaps a transparent proxy/firewall/web filter "blocked" message, or similar). Maybe try: 1. "status client=" in bconsole to check Bacula can communicate with the client. 2. If not, issue "lsof -i -P | grep 9102" at the terminal on the client, to make sure 'bacula-fd' is running (on the default port). 3. If 'bacula-fd' is listed as running, stop the Bacula File Daemon on the client to free port 9102, then run "nc -l 9102" to open a listener on the same port the file daemon uses, and send some text from the Bacula server using "nc 9102". If TCP communications are good, you should see exactly the text you type on the server appear on the Mac's terminal after pressing return. Sorry in advance if this is stuff you've already tried. Just for completeness, one of the few things I have done to the Mac in question is install Xcode (I think it replaces the shipped installation of 'make', so there's a chance it affects compilation). I'm not a big Mac user, I'm afraid. It seems that just owning a Mac automatically makes one the "Mac guy" . Thanks. -- Stephen Thompson Berkeley Seismology Lab stephen.thomp...@berkeley.edu 307 McCone Hall Office: 510.664.9177 University of California Remote: 510.214.6506 (Tue) Berkeley, CA 94720-4760 ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] Packet size too big (NOT a version mismatch)
That's a good test, which I apparently have not tried. I will do so. thanks, Stephen On 1/4/22 11:20 AM, Martin Simmons wrote: Is this happening for all backups? What happens if you run a backup with a minimal fileset that lists just one small file? __Martin On Tue, 4 Jan 2022 08:13:46 -0800, Stephen Thompson said: I am still seeing the same issue on Monterey as on Big Sur with 11.0.5 compiled from source and CoreFoundation linked in. 04-Jan 07:56 SD JobId 88: Fatal error: bsock.c:530 Packet size=1387165 too big from "client:1.2.3.4:9103". Maximum permitted 100. Terminating connection. Stephen On Tue, Jan 4, 2022 at 7:02 AM Stephen Thompson < stephen.thomp...@berkeley.edu> wrote: Graham, Thanks for presenting Monterey as a possibility! I am seeing the same issue under Monterrey as I have under Big Sur, but to know someone else does not means that it's possible. I should double check that I am using a freshly compiled client on Monterey and not just the one that I compiled on Big Sur. I am backing up Macs with bacula, but not really for system recovery, more to backup user files/documents that they may not be backing up themselves. I do note a number of Mac system files that refuse to be backed up, but again for my purposes, I do not care too much. It would be nice to be able to BMR a Mac, but not a requirement where I am at, being operationally a Linux shop. Stephen On Tue, Jan 4, 2022 at 6:20 AM Graham Sparks wrote: Hi David, I use Time Machine (for the System disk) as well as Bacula on my Mac, as I'd still need the Time Machine backup to do a bare-metal restore (with Apps). I use Bacula to back up this and an external data drive. Rather than purchasing a separate "Time Capsule", I set up Samba on a Linux VM to expose an SMB share that the Mac sees as a Time Capsule drive ( https://wiki.samba.org/index.php/Configure_Samba_to_Work_Better_with_Mac_OS_X ). I had one problem with Time Machine a few months ago, where it stopped backing up data and insisted on starting the backup 'chain' from scratch again. I was a little miffed . I'm afraid I can only confirm that the Bacula v9.6 and v11 file daemons worked for me under macOS Catalina and Monetery (I skipped Big Sur. Not for good reason---just laziness). Both v9 and v11 clients were compiled from source (setting the linker flags to "-framework CoreFoundation" as already suggested). I've personally not run in to problems with System Integrity Protection, although I do give the bacula-fd executable "Full Disk" permissions. Thanks. -- Graham Sparks From: David Brodbeck Sent: 03 January 2022 18:36 Cc: bacula-users@lists.sourceforge.net < bacula-users@lists.sourceforge.net> Subject: Re: [Bacula-users] Packet size too big (NOT a version mismatch) I'm curious if anyone has moved away from Bacula on macOS and what alternatives they're using. Even before this, it was getting more and more awkward to set up -- bacula really doesn't play well with SIP, for example, and running "csrutil disable" on every system is not a security best practice. On Wed, Dec 8, 2021 at 4:46 PM Stephen Thompson < stephen.thomp...@berkeley.edu> wrote: Disappointing... I am having the same issue on BigSur with the 11.0.5 release as I had with 9x. 08-Dec 15:42 SD JobId 878266: Fatal error: bsock.c:530 Packet size=1387166 too big from "client:1.2.3.4:8103". Maximum permitted 100. Terminating connection. Setting 'Maximum Network Buffer Size' does not appear to solve issue. Are there users out there successfully running a bacula client on Big Sur?? Stephen On Wed, Dec 1, 2021 at 3:25 PM Stephen Thompson < stephen.thomp...@berkeley.edu> wrote: Not sure if this is correct, but I've been able to at least compile bacula client 11.0.5 on Big Sur by doing before configure step: LDFLAGS='-framework CoreFoundation' We'll see next up whether it runs and whether it exhibits the issue seen under Big Sur for 9x client. Stephen On Tue, Nov 23, 2021 at 7:32 AM Stephen Thompson < stephen.thomp...@berkeley.edu> wrote: Josh, Thanks for the tip. That did not appear to be the cause of this issue, though perhaps it will fix a yet to be found issue that I would have run into after I get past this compilation error. Stephen On Mon, Nov 22, 2021 at 9:22 AM Josh Fisher wrote: On 11/22/21 10:46, Stephen Thompson wrote: All, I too was having the issue with running a 9x client on Big Sur. I've tried compiling 11.0.5 but have not found my way past: This might be due to a libtool.m4 bug having to do with MacOS changing the major Darwin version from 19.x to 20.x. There is a patch at https://www.mail-archive.com/libtool-patches@gnu.org/msg07396.html Linking bacula-fd ... /Users/bacula/src/bacula-11.0.5-CLIENT.MAC/libtool --silent --tag=CXX --mode=link /usr/bin/g++ -L../lib -L../findlib -o bacula-fd filed.o authenticate.o backup.o crypto.o win
Re: [Bacula-users] Packet size too big (NOT a version mismatch)
I am still seeing the same issue on Monterey as on Big Sur with 11.0.5 compiled from source and CoreFoundation linked in. 04-Jan 07:56 SD JobId 88: Fatal error: bsock.c:530 Packet size=1387165 too big from "client:1.2.3.4:9103". Maximum permitted 100. Terminating connection. Stephen On Tue, Jan 4, 2022 at 7:02 AM Stephen Thompson < stephen.thomp...@berkeley.edu> wrote: > > Graham, > > Thanks for presenting Monterey as a possibility! I am seeing the same > issue under Monterrey as I have under Big Sur, but to know someone else > does not means that it's possible. I should double check that I am using a > freshly compiled client on Monterey and not just the one that I compiled on > Big Sur. > > I am backing up Macs with bacula, but not really for system recovery, more > to backup user files/documents that they may not be backing up themselves. > I do note a number of Mac system files that refuse to be backed up, but > again for my purposes, I do not care too much. It would be nice to be able > to BMR a Mac, but not a requirement where I am at, being operationally a > Linux shop. > > Stephen > > > > > On Tue, Jan 4, 2022 at 6:20 AM Graham Sparks wrote: > >> Hi David, >> >> I use Time Machine (for the System disk) as well as Bacula on my Mac, as >> I'd still need the Time Machine backup to do a bare-metal restore (with >> Apps). I use Bacula to back up this and an external data drive. >> >> Rather than purchasing a separate "Time Capsule", I set up Samba on a >> Linux VM to expose an SMB share that the Mac sees as a Time Capsule drive ( >> https://wiki.samba.org/index.php/Configure_Samba_to_Work_Better_with_Mac_OS_X >> ). >> >> I had one problem with Time Machine a few months ago, where it stopped >> backing up data and insisted on starting the backup 'chain' from scratch >> again. I was a little miffed . >> >> I'm afraid I can only confirm that the Bacula v9.6 and v11 file daemons >> worked for me under macOS Catalina and Monetery (I skipped Big Sur. Not >> for good reason---just laziness). Both v9 and v11 clients were compiled >> from source (setting the linker flags to "-framework CoreFoundation" as >> already suggested). >> >> I've personally not run in to problems with System Integrity Protection, >> although I do give the bacula-fd executable "Full Disk" permissions. >> >> Thanks. >> -- >> Graham Sparks >> >> >> >> From: David Brodbeck >> Sent: 03 January 2022 18:36 >> Cc: bacula-users@lists.sourceforge.net < >> bacula-users@lists.sourceforge.net> >> Subject: Re: [Bacula-users] Packet size too big (NOT a version mismatch) >> >> I'm curious if anyone has moved away from Bacula on macOS and what >> alternatives they're using. Even before this, it was getting more and more >> awkward to set up -- bacula really doesn't play well with SIP, for example, >> and running "csrutil disable" on every system is not a security best >> practice. >> >> On Wed, Dec 8, 2021 at 4:46 PM Stephen Thompson < >> stephen.thomp...@berkeley.edu> wrote: >> >> >> Disappointing... I am having the same issue on BigSur with the 11.0.5 >> release as I had with 9x. >> >> 08-Dec 15:42 SD JobId 878266: Fatal error: bsock.c:530 Packet >> size=1387166 too big from "client:1.2.3.4:8103". Maximum permitted >> 100. Terminating connection. >> >> >> Setting 'Maximum Network Buffer Size' does not appear to solve issue. >> Are there users out there successfully running a bacula client on Big >> Sur?? >> Stephen >> >> >> >> On Wed, Dec 1, 2021 at 3:25 PM Stephen Thompson < >> stephen.thomp...@berkeley.edu> wrote: >> >> Not sure if this is correct, but I've been able to at least compile >> bacula client 11.0.5 on Big Sur by doing before configure step: >> >> LDFLAGS='-framework CoreFoundation' >> >> We'll see next up whether it runs and whether it exhibits the issue seen >> under Big Sur for 9x client. >> >> Stephen >> >> On Tue, Nov 23, 2021 at 7:32 AM Stephen Thompson < >> stephen.thomp...@berkeley.edu> wrote: >> >> Josh, >> >> Thanks for the tip. That did not appear to be the cause of this issue, >> though perhaps it will fix a yet to be found issue that I would have run >> into after I get past this compilation error. >> >> Stephen >> >> >> >> On Mon, Nov 22, 2021 at 9:22 AM Josh Fisher wrote: >> >> On 11/22/21 10:46, Steph
Re: [Bacula-users] Packet size too big (NOT a version mismatch)
Thanks Bill, you nailed it. 04-Jan-2022 16:13:52 FD: backup.c:1356-884680 fname=/Users/USER/Library/Containers/com.apple.Safari.CacheDeleteExtension snap=/Users/USER/Library/Containers/com.apple.Safari.CacheDeleteExtension link=/Users/USER/Library/Containers/com.apple.Safari.CacheDeleteExtension/ Last file before error is thrown and job craps out. I would if that's the only file, will try to exclude and see how far the job can go. Also, note a debug level of 150 is way more than needed to troubleshoot this and I canceled attempt after trace file was 60G. level=10 was enough to log which files were being backedup as well as the error that terminated job. Stephen On 1/4/22 12:03 PM, Bill Arlofski via Bacula-users wrote: On 1/4/22 12:26, Stephen Thompson wrote: Yes, backing up a single file on my problem hosts does succeed. H... Stephen Hello Stephen, This issue looked familiar to me, so I checked internally and I think I found something. I am pretty sure that this is an issue due to the larger possible size of extended attributes that Big Sur uses. From what I can gather, this has been addressed and fixed in Bacula Enterprise, and the fix will appear in the next Bacula Community release. (no ETA that I am aware of yet, but I assume very soon) In the case I found, running the FD in debug mode, level=150 revealed there was an issue with one specific file: 8< /Users//Library/Containers/com.apple.Safari.CacheDeleteExtension 8< The temporary workaround at the time (Sept 2021) was to omit this file (or whichever file your system is working on when the job fails) from the backups. No idea if this means much, but there was also a mention made: "this seems to be related to Time Machine" Setting the FD in debug mode: * setdebug level=150 options=tc trace=1 client= Then, run the backup until it fails, and stop debugging: * setdebug level=0 trace=0 client= In /opt/bacula/working on the FD (or wherever "WorkingDirectory" is set to), there will be a *.trace file. You will be looking for the file mentioned before the error: 8< bxattr.c:310-69825 Network send error to SD. ERR=Broken pipe 8< Hope this helps. Bill -- Bill Arlofski w...@protonmail.com ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users -- Stephen Thompson Berkeley Seismology Lab stephen.thomp...@berkeley.edu 307 McCone Hall Office: 510.664.9177 University of California Remote: 510.214.6506 (Tue) Berkeley, CA 94720-4760 ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] Packet size too big (NOT a version mismatch)
Josh, Thanks for the tip. That did not appear to be the cause of this issue, though perhaps it will fix a yet to be found issue that I would have run into after I get past this compilation error. Stephen On Mon, Nov 22, 2021 at 9:22 AM Josh Fisher wrote: > > On 11/22/21 10:46, Stephen Thompson wrote: > > > All, > > I too was having the issue with running a 9x client on Big Sur. I've > tried compiling 11.0.5 but have not found my way past: > > > This might be due to a libtool.m4 bug having to do with MacOS changing the > major Darwin version from 19.x to 20.x. There is a patch at > https://www.mail-archive.com/libtool-patches@gnu.org/msg07396.html > > > > Linking bacula-fd ... > > /Users/bacula/src/bacula-11.0.5-CLIENT.MAC/libtool --silent --tag=CXX > --mode=link /usr/bin/g++ -L../lib -L../findlib -o bacula-fd filed.o > authenticate.o backup.o crypto.o win_efs.o estimate.o fdcollect.o > fd_plugins.o accurate.o bacgpfs.o filed_conf.o runres_conf.o heartbeat.o > hello.o job.o fd_snapshot.o restore.o status.o verify.o verify_vol.o > fdcallsdir.o suspend.o org_filed_dedup.o bacl.o bacl_osx.o bxattr.o > bxattr_osx.o \ > > -lz -lbacfind -lbaccfg -lbac -lm -lpthread \ > > -L/usr/local/opt/openssl@1.1/lib -lssl -lcrypto-framework IOKit > > Undefined symbols for architecture x86_64: > > "___CFConstantStringClassReference", referenced from: > > CFString in suspend.o > > CFString in suspend.o > > ld: symbol(s) not found for architecture x86_64 > > clang: error: linker command failed with exit code 1 (use -v to see > invocation) > > make[1]: *** [bacula-fd] Error 1 > > > > Seems like this might have something to do with the expection of headers > being here: > > /System/Library/Frameworks/CoreFoundation.framework/Headers > > when they are here: > > > /Library/Developer/CommandLineTools/SDKs/MacOSX11.0.sdk/System/Library/Frameworks/CoreFoundation.framework/Headers/ > but that may be a red herring. > > There also appears to be a 'clang' in two locations on OS X, /usr and > xcode subdir. Hmm > > Stephen > > On Tue, Nov 16, 2021 at 12:00 AM Eric Bollengier via Bacula-users < > bacula-users@lists.sourceforge.net> wrote: > >> Hello, >> >> On 11/15/21 21:46, David Brodbeck wrote: >> > To do that I'd have to upgrade the director and the storage first, >> right? >> > (Director can't be an earlier version than the FD, and the SD must have >> the >> > same version as the director.) >> >> In general yes, the code is designed to support Old FDs but can have >> problems >> with newer FDs. In your case it may work. >> >> At least, you can try a status client to see if the problem is solved and >> if you can run a backup & a restore. >> >> Best Regards, >> Eric >> >> >> ___ >> Bacula-users mailing list >> Bacula-users@lists.sourceforge.net >> https://lists.sourceforge.net/lists/listinfo/bacula-users >> > > > -- > Stephen Thompson Berkeley Seismology Lab > stephen.thomp...@berkeley.edu 307 McCone Hall > Office: 510.664.9177 University of California > Remote: 510.214.6506 (Tue) Berkeley, CA 94720-4760 > > > ___ > Bacula-users mailing > listBacula-users@lists.sourceforge.nethttps://lists.sourceforge.net/lists/listinfo/bacula-users > > -- Stephen Thompson Berkeley Seismology Lab stephen.thomp...@berkeley.edu 307 McCone Hall Office: 510.664.9177 University of California Remote: 510.214.6506 (Tue) Berkeley, CA 94720-4760 ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] Packet size too big (NOT a version mismatch)
All, I too was having the issue with running a 9x client on Big Sur. I've tried compiling 11.0.5 but have not found my way past: Linking bacula-fd ... /Users/bacula/src/bacula-11.0.5-CLIENT.MAC/libtool --silent --tag=CXX --mode=link /usr/bin/g++ -L../lib -L../findlib -o bacula-fd filed.o authenticate.o backup.o crypto.o win_efs.o estimate.o fdcollect.o fd_plugins.o accurate.o bacgpfs.o filed_conf.o runres_conf.o heartbeat.o hello.o job.o fd_snapshot.o restore.o status.o verify.o verify_vol.o fdcallsdir.o suspend.o org_filed_dedup.o bacl.o bacl_osx.o bxattr.o bxattr_osx.o \ -lz -lbacfind -lbaccfg -lbac -lm -lpthread \ -L/usr/local/opt/openssl@1.1/lib -lssl -lcrypto-framework IOKit Undefined symbols for architecture x86_64: "___CFConstantStringClassReference", referenced from: CFString in suspend.o CFString in suspend.o ld: symbol(s) not found for architecture x86_64 clang: error: linker command failed with exit code 1 (use -v to see invocation) make[1]: *** [bacula-fd] Error 1 Seems like this might have something to do with the expection of headers being here: /System/Library/Frameworks/CoreFoundation.framework/Headers when they are here: /Library/Developer/CommandLineTools/SDKs/MacOSX11.0.sdk/System/Library/Frameworks/CoreFoundation.framework/Headers/ but that may be a red herring. There also appears to be a 'clang' in two locations on OS X, /usr and xcode subdir. Hmm Stephen On Tue, Nov 16, 2021 at 12:00 AM Eric Bollengier via Bacula-users < bacula-users@lists.sourceforge.net> wrote: > Hello, > > On 11/15/21 21:46, David Brodbeck wrote: > > To do that I'd have to upgrade the director and the storage first, right? > > (Director can't be an earlier version than the FD, and the SD must have > the > > same version as the director.) > > In general yes, the code is designed to support Old FDs but can have > problems > with newer FDs. In your case it may work. > > At least, you can try a status client to see if the problem is solved and > if you can run a backup & a restore. > > Best Regards, > Eric > > > ___ > Bacula-users mailing list > Bacula-users@lists.sourceforge.net > https://lists.sourceforge.net/lists/listinfo/bacula-users > -- Stephen Thompson Berkeley Seismology Lab stephen.thomp...@berkeley.edu 307 McCone Hall Office: 510.664.9177 University of California Remote: 510.214.6506 (Tue) Berkeley, CA 94720-4760 ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] Packet size too big (NOT a version mismatch)
David, Sorry I can't offer a solution, but I can report that am I getting the same error when trying to run bacula-fd 9.x on Big Sur (hand compiled). I've tried the other suggestion of Maximum Network Buffer Size to no avail. Stephen On 11/12/21 2:14 PM, David Brodbeck wrote: I'm getting this error trying to back up a macOS client. I recently re-installed bacula from macports on this client, after an upgrade to macOS Big Sur. | russell.math.ucsb.edu-sd JobId 80985: Fatal error: bsock.c:520 Packet size=1387166 too big from "client:128.111.88.29:62571 <http://128.111.88.29:62571>". Maximum permitted 100. Terminating connection. | Normally when I've seen this it's because of a version mismatch between the client and the director or storage daemon, but that's not the case here; the director, sd, and fd are all running the same version: 1000 OK: 103 self-help.math.ucsb.edu-dir Version: 9.4.4 (28 May 2019) russell.math.ucsb.edu-sd Version: 9.4.4 (28 May 2019) x86_64-pc-linux-gnu redhat (Core) noether.math.ucsb.edu-fd Version: 9.4.4 (28 May 2019) x86_64-apple-darwin20.6.0 osx 20.6.0 All except the fd are built directly from bacula source. (The fd was built with macports.) Any suggestions on where to look? Other clients are backing up fine to the same sd, so I feel like it must be a client configuration issue, but I can't figure out how. -- David Brodbeck (they/them) System Administrator, Department of Mathematics University of California, Santa Barbara ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users -- Stephen Thompson Berkeley Seismology Lab stephen.thomp...@berkeley.edu 307 McCone Hall Office: 510.664.9177 University of California Remote: 510.214.6506 (Tue) Berkeley, CA 94720-4760 ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] Packet size too big (NOT a version mismatch)
Not sure if this is correct, but I've been able to at least compile bacula client 11.0.5 on Big Sur by doing before configure step: LDFLAGS='-framework CoreFoundation' We'll see next up whether it runs and whether it exhibits the issue seen under Big Sur for 9x client. Stephen On Tue, Nov 23, 2021 at 7:32 AM Stephen Thompson < stephen.thomp...@berkeley.edu> wrote: > > Josh, > > Thanks for the tip. That did not appear to be the cause of this issue, > though perhaps it will fix a yet to be found issue that I would have run > into after I get past this compilation error. > > Stephen > > > > On Mon, Nov 22, 2021 at 9:22 AM Josh Fisher wrote: > >> >> On 11/22/21 10:46, Stephen Thompson wrote: >> >> >> All, >> >> I too was having the issue with running a 9x client on Big Sur. I've >> tried compiling 11.0.5 but have not found my way past: >> >> >> This might be due to a libtool.m4 bug having to do with MacOS changing >> the major Darwin version from 19.x to 20.x. There is a patch at >> https://www.mail-archive.com/libtool-patches@gnu.org/msg07396.html >> >> >> >> Linking bacula-fd ... >> >> /Users/bacula/src/bacula-11.0.5-CLIENT.MAC/libtool --silent --tag=CXX >> --mode=link /usr/bin/g++ -L../lib -L../findlib -o bacula-fd filed.o >> authenticate.o backup.o crypto.o win_efs.o estimate.o fdcollect.o >> fd_plugins.o accurate.o bacgpfs.o filed_conf.o runres_conf.o heartbeat.o >> hello.o job.o fd_snapshot.o restore.o status.o verify.o verify_vol.o >> fdcallsdir.o suspend.o org_filed_dedup.o bacl.o bacl_osx.o bxattr.o >> bxattr_osx.o \ >> >> -lz -lbacfind -lbaccfg -lbac -lm -lpthread \ >> >> -L/usr/local/opt/openssl@1.1/lib -lssl -lcrypto-framework IOKit >> >> Undefined symbols for architecture x86_64: >> >> "___CFConstantStringClassReference", referenced from: >> >> CFString in suspend.o >> >> CFString in suspend.o >> >> ld: symbol(s) not found for architecture x86_64 >> >> clang: error: linker command failed with exit code 1 (use -v to see >> invocation) >> >> make[1]: *** [bacula-fd] Error 1 >> >> >> >> Seems like this might have something to do with the expection of headers >> being here: >> >> /System/Library/Frameworks/CoreFoundation.framework/Headers >> >> when they are here: >> >> >> /Library/Developer/CommandLineTools/SDKs/MacOSX11.0.sdk/System/Library/Frameworks/CoreFoundation.framework/Headers/ >> but that may be a red herring. >> >> There also appears to be a 'clang' in two locations on OS X, /usr and >> xcode subdir. Hmm >> >> Stephen >> >> On Tue, Nov 16, 2021 at 12:00 AM Eric Bollengier via Bacula-users < >> bacula-users@lists.sourceforge.net> wrote: >> >>> Hello, >>> >>> On 11/15/21 21:46, David Brodbeck wrote: >>> > To do that I'd have to upgrade the director and the storage first, >>> right? >>> > (Director can't be an earlier version than the FD, and the SD must >>> have the >>> > same version as the director.) >>> >>> In general yes, the code is designed to support Old FDs but can have >>> problems >>> with newer FDs. In your case it may work. >>> >>> At least, you can try a status client to see if the problem is solved and >>> if you can run a backup & a restore. >>> >>> Best Regards, >>> Eric >>> >>> >>> ___ >>> Bacula-users mailing list >>> Bacula-users@lists.sourceforge.net >>> https://lists.sourceforge.net/lists/listinfo/bacula-users >>> >> >> >> -- >> Stephen Thompson Berkeley Seismology Lab >> stephen.thomp...@berkeley.edu 307 McCone Hall >> Office: 510.664.9177 University of California >> Remote: 510.214.6506 (Tue) Berkeley, CA 94720-4760 >> >> >> ___ >> Bacula-users mailing >> listBacula-users@lists.sourceforge.nethttps://lists.sourceforge.net/lists/listinfo/bacula-users >> >> > > -- > Stephen Thompson Berkeley Seismology Lab > stephen.thomp...@berkeley.edu 307 McCone Hall > Office: 510.664.9177 University of California > Remote: 510.214.6506 (Tue) Berkeley, CA 94720-4760 > -- Stephen Thompson Berkeley Seismology Lab stephen.thomp...@berkeley.edu 307 McCone Hall Office: 510.664.9177 University of California Remote: 510.214.6506 (Tue) Berkeley, CA 94720-4760 ___ Bacula-users mailing list Bacula-users@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/bacula-users
Re: [Bacula-users] Packet size too big (NOT a version mismatch)
Disappointing... I am having the same issue on BigSur with the 11.0.5 release as I had with 9x. 08-Dec 15:42 SD JobId 878266: Fatal error: bsock.c:530 Packet size=1387166 too big from "client:1.2.3.4:8103". Maximum permitted 100. Terminating connection. Setting 'Maximum Network Buffer Size' does not appear to solve issue. Are there users out there successfully running a bacula client on Big Sur?? Stephen On Wed, Dec 1, 2021 at 3:25 PM Stephen Thompson < stephen.thomp...@berkeley.edu> wrote: > > Not sure if this is correct, but I've been able to at least compile bacula > client 11.0.5 on Big Sur by doing before configure step: > > LDFLAGS='-framework CoreFoundation' > > We'll see next up whether it runs and whether it exhibits the issue seen > under Big Sur for 9x client. > > Stephen > > On Tue, Nov 23, 2021 at 7:32 AM Stephen Thompson < > stephen.thomp...@berkeley.edu> wrote: > >> >> Josh, >> >> Thanks for the tip. That did not appear to be the cause of this issue, >> though perhaps it will fix a yet to be found issue that I would have run >> into after I get past this compilation error. >> >> Stephen >> >> >> >> On Mon, Nov 22, 2021 at 9:22 AM Josh Fisher wrote: >> >>> >>> On 11/22/21 10:46, Stephen Thompson wrote: >>> >>> >>> All, >>> >>> I too was having the issue with running a 9x client on Big Sur. I've >>> tried compiling 11.0.5 but have not found my way past: >>> >>> >>> This might be due to a libtool.m4 bug having to do with MacOS changing >>> the major Darwin version from 19.x to 20.x. There is a patch at >>> https://www.mail-archive.com/libtool-patches@gnu.org/msg07396.html >>> >>> >>> >>> Linking bacula-fd ... >>> >>> /Users/bacula/src/bacula-11.0.5-CLIENT.MAC/libtool --silent --tag=CXX >>> --mode=link /usr/bin/g++ -L../lib -L../findlib -o bacula-fd filed.o >>> authenticate.o backup.o crypto.o win_efs.o estimate.o fdcollect.o >>> fd_plugins.o accurate.o bacgpfs.o filed_conf.o runres_conf.o heartbeat.o >>> hello.o job.o fd_snapshot.o restore.o status.o verify.o verify_vol.o >>> fdcallsdir.o suspend.o org_filed_dedup.o bacl.o bacl_osx.o bxattr.o >>> bxattr_osx.o \ >>> >>> -lz -lbacfind -lbaccfg -lbac -lm -lpthread \ >>> >>> -L/usr/local/opt/openssl@1.1/lib -lssl -lcrypto-framework IOKit >>> >>> Undefined symbols for architecture x86_64: >>> >>> "___CFConstantStringClassReference", referenced from: >>> >>> CFString in suspend.o >>> >>> CFString in suspend.o >>> >>> ld: symbol(s) not found for architecture x86_64 >>> >>> clang: error: linker command failed with exit code 1 (use -v to see >>> invocation) >>> >>> make[1]: *** [bacula-fd] Error 1 >>> >>> >>> >>> Seems like this might have something to do with the expection of headers >>> being here: >>> >>> /System/Library/Frameworks/CoreFoundation.framework/Headers >>> >>> when they are here: >>> >>> >>> /Library/Developer/CommandLineTools/SDKs/MacOSX11.0.sdk/System/Library/Frameworks/CoreFoundation.framework/Headers/ >>> but that may be a red herring. >>> >>> There also appears to be a 'clang' in two locations on OS X, /usr and >>> xcode subdir. Hmm >>> >>> Stephen >>> >>> On Tue, Nov 16, 2021 at 12:00 AM Eric Bollengier via Bacula-users < >>> bacula-users@lists.sourceforge.net> wrote: >>> >>>> Hello, >>>> >>>> On 11/15/21 21:46, David Brodbeck wrote: >>>> > To do that I'd have to upgrade the director and the storage first, >>>> right? >>>> > (Director can't be an earlier version than the FD, and the SD must >>>> have the >>>> > same version as the director.) >>>> >>>> In general yes, the code is designed to support Old FDs but can have >>>> problems >>>> with newer FDs. In your case it may work. >>>> >>>> At least, you can try a status client to see if the problem is solved >>>> and >>>> if you can run a backup & a restore. >>>> >>>> Best Regards, >>>> Eric >>>> >>>> >>>> ___ >>>> Bacula-users mailing list >>>> Ba
Re: [Bacula-users] Packet size too big (NOT a version mismatch)
Graham, Thanks for presenting Monterey as a possibility! I am seeing the same issue under Monterrey as I have under Big Sur, but to know someone else does not means that it's possible. I should double check that I am using a freshly compiled client on Monterey and not just the one that I compiled on Big Sur. I am backing up Macs with bacula, but not really for system recovery, more to backup user files/documents that they may not be backing up themselves. I do note a number of Mac system files that refuse to be backed up, but again for my purposes, I do not care too much. It would be nice to be able to BMR a Mac, but not a requirement where I am at, being operationally a Linux shop. Stephen On Tue, Jan 4, 2022 at 6:20 AM Graham Sparks wrote: > Hi David, > > I use Time Machine (for the System disk) as well as Bacula on my Mac, as > I'd still need the Time Machine backup to do a bare-metal restore (with > Apps). I use Bacula to back up this and an external data drive. > > Rather than purchasing a separate "Time Capsule", I set up Samba on a > Linux VM to expose an SMB share that the Mac sees as a Time Capsule drive ( > https://wiki.samba.org/index.php/Configure_Samba_to_Work_Better_with_Mac_OS_X > ). > > I had one problem with Time Machine a few months ago, where it stopped > backing up data and insisted on starting the backup 'chain' from scratch > again. I was a little miffed . > > I'm afraid I can only confirm that the Bacula v9.6 and v11 file daemons > worked for me under macOS Catalina and Monetery (I skipped Big Sur. Not > for good reason---just laziness). Both v9 and v11 clients were compiled > from source (setting the linker flags to "-framework CoreFoundation" as > already suggested). > > I've personally not run in to problems with System Integrity Protection, > although I do give the bacula-fd executable "Full Disk" permissions. > > Thanks. > -- > Graham Sparks > > > > From: David Brodbeck > Sent: 03 January 2022 18:36 > Cc: bacula-users@lists.sourceforge.net > > Subject: Re: [Bacula-users] Packet size too big (NOT a version mismatch) > > I'm curious if anyone has moved away from Bacula on macOS and what > alternatives they're using. Even before this, it was getting more and more > awkward to set up -- bacula really doesn't play well with SIP, for example, > and running "csrutil disable" on every system is not a security best > practice. > > On Wed, Dec 8, 2021 at 4:46 PM Stephen Thompson < > stephen.thomp...@berkeley.edu> wrote: > > > Disappointing... I am having the same issue on BigSur with the 11.0.5 > release as I had with 9x. > > 08-Dec 15:42 SD JobId 878266: Fatal error: bsock.c:530 Packet size=1387166 > too big from "client:1.2.3.4:8103". Maximum permitted 100. > Terminating connection. > > > Setting 'Maximum Network Buffer Size' does not appear to solve issue. > Are there users out there successfully running a bacula client on Big Sur?? > Stephen > > > > On Wed, Dec 1, 2021 at 3:25 PM Stephen Thompson < > stephen.thomp...@berkeley.edu> wrote: > > Not sure if this is correct, but I've been able to at least compile bacula > client 11.0.5 on Big Sur by doing before configure step: > > LDFLAGS='-framework CoreFoundation' > > We'll see next up whether it runs and whether it exhibits the issue seen > under Big Sur for 9x client. > > Stephen > > On Tue, Nov 23, 2021 at 7:32 AM Stephen Thompson < > stephen.thomp...@berkeley.edu> wrote: > > Josh, > > Thanks for the tip. That did not appear to be the cause of this issue, > though perhaps it will fix a yet to be found issue that I would have run > into after I get past this compilation error. > > Stephen > > > > On Mon, Nov 22, 2021 at 9:22 AM Josh Fisher wrote: > > On 11/22/21 10:46, Stephen Thompson wrote: > > All, > > I too was having the issue with running a 9x client on Big Sur. I've > tried compiling 11.0.5 but have not found my way past: > > This might be due to a libtool.m4 bug having to do with MacOS changing the > major Darwin version from 19.x to 20.x. There is a patch at > https://www.mail-archive.com/libtool-patches@gnu.org/msg07396.html > > > Linking bacula-fd ... > /Users/bacula/src/bacula-11.0.5-CLIENT.MAC/libtool --silent --tag=CXX > --mode=link /usr/bin/g++ -L../lib -L../findlib -o bacula-fd filed.o > authenticate.o backup.o crypto.o win_efs.o estimate.o fdcollect.o > fd_plugins.o accurate.o bacgpfs.o filed_conf.o runres_conf.o heartbeat.o > hello.o job.o fd_snapshot.o restore.o status.o verify.o verify_vol.o > fdcallsdir.o suspend.o org_filed_dedup.o bacl.o bacl_osx.o bxattr.o > bxattr_osx.o \ > -lz -lb
