sysctl -zarc for ZFS users
Hi all. I've been watching ZFS activites on my machine and improved systat to monitor such. One of my first goals is to watch ZFS cache statistics. I posted my patch to the bugzilla @ https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=195460 I've been using on 11-CURRENT and 10.1-RELEASE myself. Regards, Hiro % systat -zarc Total MFU MRUAnon Hdr L2Hdr Other ZFS ARC206M 63M136M826K 1102K 0K 5251K ratehits misses total hits total misses arcstats : 75%1183 38823721 4202 arcstats.demand_data : 0% 0 0 5280 arcstats.demand_metadata : 76%1007 30821441 2757 arcstats.prefetch_data: 0% 0 000 arcstats.prefetch_metadata: 68% 176 80 1752 1445 zfetchstats : 44% 95412042641025482 arcstats.l2 : 0% 0 000 vdev_cache_stats : 0% 0 000 ___ freebsd-current@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-current To unsubscribe, send any mail to freebsd-current-unsubscr...@freebsd.org
[CFT] pkg 1.4.0 rc2
Hi, We have released a new 1.4.0 rc2 version of pkg (available in ports-mgmt/pkg-devel) since first beta it has received tons of bug fixes and should be now way more reliable and able to handle ootb without mistakes upgrades like the gettext one and the perl one. All reported issues should have been fixed since. Please test that new version I would like to make it the final release if possible. Best regards, Bapt pgp5XgT9byb8u.pgp Description: PGP signature
Re: [CFT] pkg 1.4.0 rc2
On Sat, Dec 06, 2014 at 05:12:12AM -0800, Jeffrey Bouquet wrote: On 12/06/14 04:40, Baptiste Daroussin wrote: Hi, We have released a new 1.4.0 rc2 version of pkg (available in ports-mgmt/pkg-devel) since first beta it has received tons of bug fixes and should be now way more reliable and able to handle ootb without mistakes upgrades like the gettext one and the perl one. All reported issues should have been fixed since. Please test that new version I would like to make it the final release if possible. Best regards, Bapt The upgrade of pkg-devel went slightly more without problems. Did not test a direct deinstall/reinstall though. OTOH pkg install xorriso xombrero still wants to remove w3m-img , install w3m, guile 1.8, x246, along with the reinstall of those two. For which I n and pkg delete xorriso xombrero cd ../cache/pkg pkg clean pkg install xorr[tab] xomb[tab] to complete and install manually Works, but a few more commands in the way. I seem to have no guile installed (was using v2 for mcron) and libx264 installed. Just ignoring guile for now probably. Sounds like a port problem that pkg is showing I'll dig into it. regards, Bapt pgp86MPk5dm5O.pgp Description: PGP signature
Re: [CFT] pkg 1.4.0 rc2
On 12/06/14 04:40, Baptiste Daroussin wrote: Hi, We have released a new 1.4.0 rc2 version of pkg (available in ports-mgmt/pkg-devel) since first beta it has received tons of bug fixes and should be now way more reliable and able to handle ootb without mistakes upgrades like the gettext one and the perl one. All reported issues should have been fixed since. Please test that new version I would like to make it the final release if possible. Best regards, Bapt The upgrade of pkg-devel went slightly more without problems. Did not test a direct deinstall/reinstall though. OTOH pkg install xorriso xombrero still wants to remove w3m-img , install w3m, guile 1.8, x246, along with the reinstall of those two. For which I n and pkg delete xorriso xombrero cd ../cache/pkg pkg clean pkg install xorr[tab] xomb[tab] to complete and install manually Works, but a few more commands in the way. I seem to have no guile installed (was using v2 for mcron) and libx264 installed. Just ignoring guile for now probably. ___ freebsd-current@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-current To unsubscribe, send any mail to freebsd-current-unsubscr...@freebsd.org
Re: [CFT] pkg 1.4.0 rc2
On Sat, Dec 06, 2014 at 05:12:12AM -0800, Jeffrey Bouquet wrote: On 12/06/14 04:40, Baptiste Daroussin wrote: Hi, We have released a new 1.4.0 rc2 version of pkg (available in ports-mgmt/pkg-devel) since first beta it has received tons of bug fixes and should be now way more reliable and able to handle ootb without mistakes upgrades like the gettext one and the perl one. All reported issues should have been fixed since. Please test that new version I would like to make it the final release if possible. Best regards, Bapt The upgrade of pkg-devel went slightly more without problems. Did not test a direct deinstall/reinstall though. OTOH pkg install xorriso xombrero still wants to remove w3m-img , install w3m, guile 1.8, x246, along with the reinstall of those two. Right so the is a problem in the way ports express dependencies and can only be solved with some provides/requires we do not know yet how to express with the ports tree :( Regards, Bapt pgp56FNINKiTW.pgp Description: PGP signature
Re: sysctl -zarc for ZFS users
2014-12-06 11:35 GMT+01:00 Yoshihiro Ota o...@j.email.ne.jp: Hi all. I've been watching ZFS activites on my machine and improved systat to monitor such. One of my first goals is to watch ZFS cache statistics. I posted my patch to the bugzilla @ https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=195460 I've been using on 11-CURRENT and 10.1-RELEASE myself. Regards, Hiro % systat -zarc Total MFU MRUAnon Hdr L2Hdr Other ZFS ARC206M 63M136M826K 1102K 0K 5251K ratehits misses total hits total misses arcstats : 75%1183 38823721 4202 arcstats.demand_data : 0% 0 0 528 0 arcstats.demand_metadata : 76%1007 30821441 2757 arcstats.prefetch_data: 0% 0 00 0 arcstats.prefetch_metadata: 68% 176 80 1752 1445 zfetchstats : 44% 954120426410 25482 arcstats.l2 : 0% 0 00 0 vdev_cache_stats : 0% 0 00 0 Hi Hiro, just applied the patch, but I receive the error: # make make: don't know how to make zarc.c. Stop make: stopped in /usr/src/usr.bin/systat The file zarc,c is missing. Where can I find it ? Thanks. Maurizio ___ freebsd-current@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-current To unsubscribe, send any mail to freebsd-current-unsubscr...@freebsd.org
Re: Build failed in Jenkins: Build-UFS-image #599
[trimmed CC list to -current] On 06 Dec 2014, at 04:59, Garrett Cooper yaneurab...@gmail.com wrote: On Dec 5, 2014, at 16:52, jenkins-ad...@freebsd.org wrote: See https://jenkins.freebsd.org/job/Build-UFS-image/599/ I’m not entirely sure why the could not determine COMPILER_TYPE error popped up, but I have a couple of questions/concerns related to the makefile snippet. 1. Does it make sense to check CC when running make install? Yes, of course it makes sense, if parts of the install depend on e.g. COMPILER_TYPE. In some cases, you will have to run ${CC} to determine what it is, specifically if it is just cc. 2. Why isn’t this value determined once in Makefile.inc1 (per build phase), then passed down from there Because you are supposed to be able to build stuff in a subdirectory, without invoking the full top-level Makefile infrastructure. The actual infrastructure is in share/mk/bsd.*.mk, in fact. (I’ve already considered the scenario where someone explicitly sets CC in a non-toplevel Makefile, which is a problem, but an outlier rather than the norm)? AFAICT, it gets recomputed for every recursive make, which contributes to useless forking for something that honestly doesn’t change all that often/at all. This is indeed a pity, and if you know a better solution, let's hear it, please. :-) At EMC/Isilon at least, we set CC/CXX=false when running make distribute*/installkernel/installworld to catch logic errors with rebuilding code. Should this be in FreeBSD? Not sure what that is meant to achieve. If parts of the installation depend on the value of CC, why would you want to set it to false? Just so it can error out at those points? -Dimitry signature.asc Description: Message signed with OpenPGP using GPGMail
Re: Build failed in Jenkins: Build-UFS-image #599
On Sat, Dec 06, 2014 at 04:01:26PM +0100, Dimitry Andric wrote: [trimmed CC list to -current] On 06 Dec 2014, at 04:59, Garrett Cooper yaneurab...@gmail.com wrote: 2. Why isn?t this value determined once in Makefile.inc1 (per build phase), then passed down from there Because you are supposed to be able to build stuff in a subdirectory, without invoking the full top-level Makefile infrastructure. The actual infrastructure is in share/mk/bsd.*.mk, in fact. (I?ve already considered the scenario where someone explicitly sets CC in a non-toplevel Makefile, which is a problem, but an outlier rather than the norm)? AFAICT, it gets recomputed for every recursive make, which contributes to useless forking for something that honestly doesn?t change all that often/at all. This is indeed a pity, and if you know a better solution, let's hear it, please. :-) Why not also put this information into some environment variables with known ugly names, in top-level. Then, the lower-level calls of infrastructure first check the vars, and recalculate the compiler specific info if vars are absent. Use information from vars if present. ___ freebsd-current@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-current To unsubscribe, send any mail to freebsd-current-unsubscr...@freebsd.org
Re: Build failed in Jenkins: Build-UFS-image #599
On Sat, 2014-12-06 at 16:01 +0100, Dimitry Andric wrote: [trimmed CC list to -current] On 06 Dec 2014, at 04:59, Garrett Cooper yaneurab...@gmail.com wrote: On Dec 5, 2014, at 16:52, jenkins-ad...@freebsd.org wrote: See https://jenkins.freebsd.org/job/Build-UFS-image/599/ I’m not entirely sure why the could not determine COMPILER_TYPE error popped up, but I have a couple of questions/concerns related to the makefile snippet. 1. Does it make sense to check CC when running make install? Yes, of course it makes sense, if parts of the install depend on e.g. COMPILER_TYPE. In some cases, you will have to run ${CC} to determine what it is, specifically if it is just cc. 2. Why isn’t this value determined once in Makefile.inc1 (per build phase), then passed down from there Because you are supposed to be able to build stuff in a subdirectory, without invoking the full top-level Makefile infrastructure. The actual infrastructure is in share/mk/bsd.*.mk, in fact. (I’ve already considered the scenario where someone explicitly sets CC in a non-toplevel Makefile, which is a problem, but an outlier rather than the norm)? AFAICT, it gets recomputed for every recursive make, which contributes to useless forking for something that honestly doesn’t change all that often/at all. This is indeed a pity, and if you know a better solution, let's hear it, please. :-) At EMC/Isilon at least, we set CC/CXX=false when running make distribute*/installkernel/installworld to catch logic errors with rebuilding code. Should this be in FreeBSD? Not sure what that is meant to achieve. If parts of the installation depend on the value of CC, why would you want to set it to false? Just so it can error out at those points? -Dimitry I suspect it's to prevent building during the install phase, because doing so is astonishing. I was astonished the other day when it happened. I would be much happier if it had just errored out. -- Ian ___ freebsd-current@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-current To unsubscribe, send any mail to freebsd-current-unsubscr...@freebsd.org
Re: [CFT] pkg 1.4.0 rc2
Hi, Just a simple question: Does pkg upgrade automatically lock packages which are not available or cannot be built? I recently did a pkg upgrade and found myself without a working libreoffice ... 9-stable amd64 --HPS ___ freebsd-current@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-current To unsubscribe, send any mail to freebsd-current-unsubscr...@freebsd.org
Re: [CFT] pkg 1.4.0 rc2
On Sat, Dec 06, 2014 at 04:15:45PM +0100, Marc UBM wrote: On Sat, 6 Dec 2014 13:40:29 +0100 Baptiste Daroussin b...@freebsd.org wrote: Hi, We have released a new 1.4.0 rc2 version of pkg (available in ports-mgmt/pkg-devel) since first beta it has received tons of bug fixes and should be now way more reliable and able to handle ootb without mistakes upgrades like the gettext one and the perl one. All reported issues should have been fixed since. Please test that new version I would like to make it the final release if possible. Best regards, Bapt Unfortunately, installing / upgrading pkg-devel still fails with: Compressing man pages (compress-man) === Installing for pkg-1.4.0.b2 === Checking if pkg already installed pkg-static: sqlite error while executing DROP INDEX packages_unique;CREATE UNIQUE INDEX packages_unique ON packages(name); in file pkgdb.c:2245: UNIQUE constraint failed: packages.name *** Error code 74 Stop. make[1]: stopped in /usr/ports/ports-mgmt/pkg-devel *** Error code 1 Stop. make: stopped in /usr/ports/ports-mgmt/pkg-devel Any new ideas? Bye Marc ___ freebsd-current@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-current To unsubscribe, send any mail to freebsd-current-unsubscr...@freebsd.org You are trying to install beta2 not rc2 Plus be aware that your setup has outdated packages as you should not anymore have duplicated package names for around a year, but pkg rc2 will be able to handle it anyway :) regards, Bapt pgpxkViA_k0ef.pgp Description: PGP signature
Re: [CFT] pkg 1.4.0 rc2
On Sat, Dec 06, 2014 at 06:23:21PM +0100, Hans Petter Selasky wrote: Hi, Just a simple question: Does pkg upgrade automatically lock packages which are not available or cannot be built? I recently did a pkg upgrade and found myself without a working libreoffice ... 9-stable amd64 Nope it does not and will be very very hard to do it properly, the problem was that some libraries are now dropping (upstream) support for gcc 4.2 which resulted in libreoffice not being built, he said libraries was reverted to the ancient version so next package build should have libreoffice again. regards, Bapt pgpYZ7_N4NeOx.pgp Description: PGP signature
Re: Xfce crashes FreeBSD 11-CURRENT
2014-11-25 21:57 GMT+01:00 Ranjan1018 . 21474...@gmail.com: Right-click on desktop - Desktop Settings… - click on the “Create a new list, or load an existing one” - Type a non existent file name in the windows with the title “Crete/Load Backdrop List” - Open The laptop freezes and displays only a blank screen with the mouse cursor and the text cursor on the top left corner. The only things I can do is to press the power button for some seconds to switch off the laptop. Note: I have some images in a Image list as background. I am running vt(4) with KMS on Intel driver. $ uname -a FreeBSD ativ 11.0-CURRENT FreeBSD 11.0-CURRENT #0 r274845: Sat Nov 22 14:59:02 CET 2014 root@ativ:/usr/obj/usr/src/sys/NEWCONS amd64 This is only a bug in xfce ! I have upgraded the laptop to r275515 and with the above procedure these two programs crashes with a core dump: xfdesktop and xfdesktop-settings. More details: The above procedure put two invalid values in Xconf, preventing xfdesktop and xfdesktop-settings to run. The commands: xfconf-query -c xfce4-desktop -p /backdrop/screen0/monitor0/last-image-list -s ~/.config/xfce4/desktop/backdrop.list xfconf-query -c xfce4-desktop -p /backdrop/screen0/monitor0/image-path -s ~/.config/xfce4/desktop/backdrop.list restores the correct values. Maurizio ___ freebsd-current@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-current To unsubscribe, send any mail to freebsd-current-unsubscr...@freebsd.org
Re: [CFT] pkg 1.4.0 rc2
On 6-12-2014 13:40, Baptiste Daroussin wrote: Hi, We have released a new 1.4.0 rc2 version of pkg (available in ports-mgmt/pkg-devel) since first beta it has received tons of bug fixes and should be now way more reliable and able to handle ootb without mistakes upgrades like the gettext one and the perl one. All reported issues should have been fixed since. Please test that new version I would like to make it the final release if possible. I missed the previous announcement, but this is bothering me for some time already I'm using this simple cronic script to reduce traffic from scripts with nothing serious to report. But pkg still triggers: - # sudo cronic pkg audit -F Cronic detected failure or error output for the command: pkg audit -F RESULT CODE: 0 ERROR OUTPUT: pkg: vulnxml file up-to-date -- And I wonder why this informational message is reported on STDERR, and other real problems on STDOUT... -- sudo cronic pkg audit Cronic detected failure or error output for the command: pkg audit RESULT CODE: 1 ERROR OUTPUT: STANDARD OUTPUT: phpMyAdmin-4.2.13 is vulnerable: phpMyAdmin -- XSS and DoS vulnerabilities CVE: CVE-2014-9219 CVE: CVE-2014-9218 WWW: http://portaudit.FreeBSD.org/c9c46fbf-7b83-11e4-a96e-6805ca0b3d42.html 1 problem(s) in the installed packages found. -- I would atleast to have case 1) also write to STDOUT. And perhaps have case 2) report on STDERR. --WjW ___ freebsd-current@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-current To unsubscribe, send any mail to freebsd-current-unsubscr...@freebsd.org