Updated ocfs2-tools-1.2.6 ebuild with page.h patch and correctly looks up the lib-folder. Reproducible: Always Steps to Reproduce:
Created attachment 131462 [details] the ebuild
<snip> �����F��[�s�F�ί�_�K{ʗH�[c9Q$��DW���ٚ*U_ ���%+�߾_7@��e+Q9����"���w~� </snip> ??? Please, attach a plaintext diff against current ebuild.
Created attachment 131483 [details] a patch for the ocfs2-tools folders then..
Created attachment 131484 [details, diff] a patch for the ocfs2-tools folders then..
attached a patch for the ebuild. Have not dug into the pages to see if there are any guides on how to submit ebuild, except that they point to this pace.. Maybe an idea could be to display some type of simple howto if someone selects to submit a new ebuild just to make it clear on what format you might want it.
Sigh... I meant running `diff -Nau ocfs2-tools-1.2.1.ebuild ocfs2-tools-1.2.6.ebuild` - not so hard? :)
*** Bug 170760 has been marked as a duplicate of this bug. ***
(In reply to comment #6) > Sigh... I meant running `diff -Nau ocfs2-tools-1.2.1.ebuild > ocfs2-tools-1.2.6.ebuild` - not so hard? :) > First time i submit a ebuild so the procedure is new to me :D submitting: diff -Nau ocfs2-tools-1.2.1.ebuild ocfs2-tools-1.2.6.ebuild page-h.patch
Created attachment 131555 [details, diff] diff of the ebuilds
Created attachment 131556 [details] the asm/page.h removal patch.
I tested the latest ocfs-1.2.6 and page-h.patch, those all work well, so request them into mainline.
OCFS2-Tools is now up to 1.3.9 (http://oss.oracle.com/projects/ocfs2-tools/files/source/v1.3/) Though there is also version 1.2.7. The ebuild should be update to reflect this, as 1.3.9 comes with many updates allowing access to new features included in the ocfs2 mainline kernel codebase.
The 1.4-prerelease stuff wasn't posted on their announce list, but I found it on the devel list -- it looks safe to add since it's backwards-compatible, and they suggested sticking with 1.2 only for enterprise distros.
Diff for 1.2.7. Added more manpages, original o2cb.init and o2cb.sysconfig moved to doc dir (or may be remove it?). page.h patch not needed.
Created attachment 140464 [details, diff] 1.2.1 - 1.2.7 diff for ebuild file
for 1.3.9 only rename of 1.2.7 ebuild needed. 1.2.7 and 1.3.9 was successfully build on amd64 by me.
Created attachment 140548 [details, diff] 1.2.1 - 1.2.7 diff for ebuild file Also removed doman (manpages already exists in image dir)
Created attachment 140632 [details] ocfs2-tools-1.3.9 full ebuild The latest patch failed for me. Also, I extended the hack at the end to put /include under /usr/.
What's failed? And `mv "${D}"/lib "${D}"/usr' will fail on 64bit system
(In reply to comment #19) > What's failed? > And `mv "${D}"/lib "${D}"/usr' will fail on 64bit system > I got this output from patch, so I just applied it by hand. I think I probably needed to apply ocfs.diff first. Anyways, I thought it better to send the whole ebuild this time around in case others ran into that as well. I can put another one up with the amd64 stuff if you like - it's just missing the keyword, inherit eutils, and the $(get_libdir) section. # patch ocfs2-tools-1.2.1.ebuild diff_1.2.1_1.2.7.ebuild.patch patching file ocfs2-tools-1.2.1.ebuild Hunk #1 FAILED at 36. Hunk #2 FAILED at 61. 2 out of 2 hunks FAILED -- saving rejects to file ocfs2-tools-1.2.1.ebuild.rej
Created attachment 140799 [details, diff] 1.2.1 - 1.2.7 diff for ebuild file fix spaces for apply patch
Created attachment 140801 [details] Ebuild for ocfs2 tools v.1.3.9 and v.1.2.7
I tried the ocfs2-tools-1.39.ebuild, but it does not work. [snip] >>> Completed installing ocfs2-tools-1.3.9 into /var/tmp/portage/sys-fs/ocfs2-tools-1.3.9/image/ ecompressdir: bzip2 -9 /usr/share/man strip: i686-pc-linux-gnu-strip --strip-unneeded -R .comment usr/sbin/ocfs2cdsl sbin/mount.ocfs2 sbin/ocfs2_hb_ctl sbin/o2cb_ctl sbin/debugfs.ocfs2 sbin/tunefs.ocfs2 sbin/mounted.ocfs2 sbin/mkfs.ocfs2 sbin/fsck.ocfs2 lib/libocfs2.a lib/libo2cb.a lib/libo2dlm.a removing executable bit: lib/libo2cb.a removing executable bit: lib/libo2dlm.a removing executable bit: lib/libocfs2.a * QA Notice: Excessive files found in the / partition * /var/tmp/portage/sys-fs/ocfs2-tools-1.3.9/image/lib/libo2cb.a /var/tmp/portage/sys-fs/ocfs2-tools-1.3.9/image/lib/libo2dlm.a /var/tmp/portage/sys-fs/ocfs2-tools-1.3.9/image/lib/libocfs2.a * * ERROR: sys-fs/ocfs2-tools-1.3.9 failed. * Call stack: * misc-functions.sh, line 609: Called install_qa_check * misc-functions.sh, line 268: Called die * The specific snippet of code: * die "static archives (*.a) and libtool library files (*.la) do not belong in /" * The die message: * static archives (*.a) and libtool library files (*.la) do not belong in / The src_install() contains make instead of emake, a FIXME ..., is this all needed? emerge --info Portage 2.1.4.4 (default-linux/x86/dev/2007.1, gcc-4.2.3, glibc-2.7-r1, 2.6.25-rc3-via i686) ================================================================= System uname: 2.6.25-rc3-via i686 VIA C7 Processor 1800MHz Timestamp of tree: Fri, 29 Feb 2008 14:00:01 +0000 app-shells/bash: 3.2_p17-r1 dev-lang/python: 2.4.4-r6 dev-python/pycrypto: 2.0.1-r6 sys-apps/baselayout: 2.0.0 sys-apps/sandbox: 1.2.18.1-r2 sys-devel/autoconf: 2.61-r1 sys-devel/automake: 1.10.1 sys-devel/binutils: 2.18-r1 sys-devel/gcc-config: 1.4.0-r4 sys-devel/libtool: 1.5.26 virtual/os-headers: 2.6.24 ACCEPT_KEYWORDS="x86" CBUILD="i686-pc-linux-gnu" CFLAGS="-O2 -march=native -fomit-frame-pointer -pipe" CHOST="i686-pc-linux-gnu" CONFIG_PROTECT="/etc" CONFIG_PROTECT_MASK="/etc/env.d /etc/gconf /etc/gentoo-release /etc/revdep-rebuild /etc/terminfo" CXXFLAGS="-O2 -march=native -fomit-frame-pointer -pipe" DISTDIR="/usr/portage/distfiles" FEATURES="distlocks metadata-transfer noinfo parallel-fetch sandbox sfperms strict unmerge-orphans userfetch" GENTOO_MIRRORS="http://mirrors.sec.informatik.tu-darmstadt.de/gentoo/" LDFLAGS="-Wl,-O1" MAKEOPTS="-s -j2" PKGDIR="/usr/portage/packages" PORTAGE_RSYNC_OPTS="--recursive --links --safe-links --perms --times --compress --force --whole-file --delete --stats --timeout=180 --exclude=/distfiles --exclude=/local --exclude=/packages" PORTAGE_TMPDIR="/var/tmp" PORTDIR="/usr/portage" PORTDIR_OVERLAY="/usr/portage/local/layman/nouveau /usr/portage/local/layman/xen /usr/portage/local/layman/openrc /usr/portage/local" USE="ipv6 minimal pam readline unicode x86 zlib" ALSA_CARDS="ali5451 als4000 atiixp atiixp-modem bt87x ca0106 cmipci emu10k1 emu10k1x ens1370 ens1371 es1938 es1968 fm801 hda-intel intel8x0 intel8x0m maestro3 trident usb-audio via82xx via82xx-modem ymfpci" ALSA_PCM_PLUGINS="adpcm alaw asym copy dmix dshare dsnoop empty extplug file hooks iec958 ioplug ladspa lfloat linear meter mulaw multi null plug rate route share shm softvol" APACHE2_MODULES="actions alias auth_basic authn_alias authn_anon authn_dbm authn_default authn_file authz_dbm authz_default authz_groupfile authz_host authz_owner authz_user autoindex cache dav dav_fs dav_lock deflate dir disk_cache env expires ext_filter file_cache filter headers include info log_config logio mem_cache mime mime_magic negotiation rewrite setenvif speling status unique_id userdir usertrack vhost_alias" ELIBC="glibc" INPUT_DEVICES="keyboard mouse evdev" KERNEL="linux" LCD_DEVICES="bayrad cfontz cfontz633 glk hd44780 lb216 lcdm001 mtxorb ncurses text" USERLAND="GNU" VIDEO_CARDS="apm ark chips cirrus cyrix dummy fbdev glint i128 i740 i810 imstt mach64 mga neomagic nsc nv r128 radeon rendition s3 s3virge savage siliconmotion sis sisusb tdfx tga trident tseng v4l vesa vga via vmware voodoo" Unset: CPPFLAGS, CTARGET, EMERGE_DEFAULT_OPTS, INSTALL_MASK, LANG, LC_ALL, LINGUAS, PORTAGE_COMPRESS, PORTAGE_COMPRESS_FLAGS, PORTAGE_RSYNC_EXTRA_OPTS
I too have a bug. Here is what I have done... (just to be sure that everything is okay) My overlay is in /usr/local/portage and I added the ocfs2-tools-1.3.9 ~x86 in the package.keywords... -BEGIN----------------------------------------- mkdir -p /usr/local/portage/sys-fs/ocfs2-tools cd /usr/local/portage/sys-fs/ocfs2-tools wget http://bugs.gentoo.org/attachment.cgi?id=140801 mv att* ocfs2-tools-1.3.9.ebuild ebuild ocfs2-tools-1.3.9.ebuild digest update-eix emerge -avt ocfs2-tools [...] /bin/sh ./mkinstalldirs /var/tmp/portage/sys-fs/ocfs2-tools-1.3.9/image///lib/pkgconfig mkdir -p -- /var/tmp/portage/sys-fs/ocfs2-tools-1.3.9/image///lib/pkgconfig for p in o2cb.pc o2dlm.pc ocfs2.pc; do \ /usr/bin/install -c -m 644 $p /var/tmp/portage/sys-fs/ocfs2-tools-1.3.9/image///lib/pkgconfig/$p; \ done dodoc: /usr/local/portage/sys-fs/ocfs2-tools/files/INSTALL.GENTOO does not exist cp: cannot stat `/usr/local/portage/sys-fs/ocfs2-tools/files/ocfs2.init': No such file or directory cp: cannot stat `/usr/local/portage/sys-fs/ocfs2-tools/files/ocfs2.conf': No such file or directory /var/tmp/portage/sys-fs/ocfs2-tools-1.3.9/temp/environment: line 98: get_libdir: command not found mv: cannot move `/var/tmp/portage/sys-fs/ocfs2-tools-1.3.9/image//' to a subdirectory of itself, `/var/tmp/portage/sys-fs/ocfs2-tools-1.3.9/image//usr/image' >>> Completed installing ocfs2-tools-1.3.9 into /var/tmp/portage/sys-fs/ocfs2-tools-1.3.9/image/ ecompressdir: bzip2 -9 /usr/share/man strip: i686-pc-linux-gnu-strip --strip-unneeded -R .comment lib/python2.4/site-packages/ocfs2interface/plistmodule.so lib/python2.4/site-packages/ocfs2interface/ocfs2module.so lib/python2.4/site-packages/ocfs2interface/o2cbmodule.so lib/python2.4/site-packages/ocfs2interface/gidlemodule.so usr/sbin/ocfs2cdsl sbin/tunefs.ocfs2 sbin/o2cb_ctl sbin/debugfs.ocfs2 sbin/mount.ocfs2 sbin/mounted.ocfs2 sbin/ocfs2_hb_ctl sbin/fsck.ocfs2 sbin/mkfs.ocfs2 lib/libo2cb.a lib/libocfs2.a lib/libo2dlm.a removing executable bit: lib/libo2cb.a removing executable bit: lib/libo2dlm.a removing executable bit: lib/libocfs2.a * QA Notice: Excessive files found in the / partition * /var/tmp/portage/sys-fs/ocfs2-tools-1.3.9/image/lib/libo2cb.a /var/tmp/portage/sys-fs/ocfs2-tools-1.3.9/image/lib/libo2dlm.a /var/tmp/portage/sys-fs/ocfs2-tools-1.3.9/image/lib/libocfs2.a * * ERROR: sys-fs/ocfs2-tools-1.3.9 failed. * Call stack: * misc-functions.sh, line 609: Called install_qa_check * misc-functions.sh, line 268: Called die * The specific snippet of code: * die "static archives (*.a) and libtool library files (*.la) do not belong in /" * The die message: * static archives (*.a) and libtool library files (*.la) do not belong in / * * If you need support, post the topmost build error, and the call stack if relevant. * A complete build log is located at '/var/tmp/portage/sys-fs/ocfs2-tools-1.3.9/temp/build.log'. * The ebuild environment file is located at '/var/tmp/portage/sys-fs/ocfs2-tools-1.3.9/temp/environment'. * This ebuild is from an overlay: '/usr/local/portage/' * !!! install_qa_check failed; exiting. * Messages for package sys-fs/ocfs2-tools-1.3.9: * * ERROR: sys-fs/ocfs2-tools-1.3.9 failed. * Call stack: * misc-functions.sh, line 609: Called install_qa_check * misc-functions.sh, line 268: Called die * The specific snippet of code: * die "static archives (*.a) and libtool library files (*.la) do not belong in /" * The die message: * static archives (*.a) and libtool library files (*.la) do not belong in / * * If you need support, post the topmost build error, and the call stack if relevant. * A complete build log is located at '/var/tmp/portage/sys-fs/ocfs2-tools-1.3.9/temp/build.log'. * The ebuild environment file is located at '/var/tmp/portage/sys-fs/ocfs2-tools-1.3.9/temp/environment'. * This ebuild is from an overlay: '/usr/local/portage/' * -----------------------------------------FIN- # emerge --info Portage 2.1.4.4 (default-linux/x86/2007.0, gcc-4.1.2, glibc-2.6.1-r0, 2.6.23-gentoo-r8 i686) ================================================================= System uname: 2.6.23-gentoo-r8 i686 Intel(R) Xeon(TM) CPU 3.60GHz Timestamp of tree: Wed, 05 Mar 2008 19:30:01 +0000 ccache version 2.4 [disabled] app-shells/bash: 3.2_p17-r1 dev-lang/python: 2.4.4-r6 dev-python/pycrypto: 2.0.1-r6 dev-util/ccache: 2.4-r7 sys-apps/baselayout: 1.12.10-r5 sys-apps/sandbox: 1.2.18.1-r2 sys-devel/autoconf: 2.61-r1 sys-devel/automake: 1.7.9-r1, 1.8.5-r3, 1.10 sys-devel/binutils: 2.18-r1 sys-devel/gcc-config: 1.4.0-r4 sys-devel/libtool: 1.5.24 virtual/os-headers: 2.6.23-r3 ACCEPT_KEYWORDS="x86" CBUILD="i686-pc-linux-gnu" CFLAGS="-O2 -march=pentium4 -mtune=pentium4 -ftracer -fomit-frame-pointer -pipe -fforce-addr -fprefetch-loop-arrays" CHOST="i686-pc-linux-gnu" CONFIG_PROTECT="/etc" CONFIG_PROTECT_MASK="/etc/env.d /etc/fonts/fonts.conf /etc/gconf /etc/php/apache2-php5/ext-active/ /etc/php/cgi-php5/ext-active/ /etc/php/cli-php5/ext-active/ /etc/revdep-rebuild /etc/terminfo /etc/udev/rules.d" CXXFLAGS="-O2 -march=pentium4 -mtune=pentium4 -ftracer -fomit-frame-pointer -pipe -fforce-addr -fprefetch-loop-arrays" DISTDIR="/usr/portage/distfiles" FEATURES="distlocks metadata-transfer sandbox sfperms strict unmerge-orphans userfetch" GENTOO_MIRRORS="http://distfiles.gentoo.org/ http://adelie.polymtl.ca/ ftp://distro.ibiblio.org/pub/linux/distributions/gentoo/ http://distro.ibiblio.org/pub/linux/distributions/gentoo/ ftp://ftp.gtlib.gatech.edu/pub/gentoo http://www.gtlib.gatech.edu/pub/gentoo ftp://mirror.iawnet.sandia.gov/pub/gentoo/ ftp://ftp.ussg.iu.edu/pub/linux/gentoo ftp://ftp.ucsb.edu/pub/mirrors/linux/gentoo/ http://ftp.ucsb.edu/pub/mirrors/linux/gentoo/ http://gentoo.seren.com/gentoo http://gentoo.chem.wisc.edu/gentoo/ ftp://gentoo.chem.wisc.edu/gentoo/ http://cudlug.cudenver.edu/gentoo/ ftp://cudlug.cudenver.edu/pub/mirrors/distributions/gentoo/ http://gentoo.mirrors.pair.com/ ftp://gentoo.mirrors.pair.com/ http://gentoo.mirrors.tds.net/gentoo http://gentoo.netnitco.net ftp://gentoo.arcticnetwork.ca/pub/gentoo/ http://gentoo.arcticnetwork.ca/" MAKEOPTS="-j2" PKGDIR="/usr/portage/packages" PORTAGE_RSYNC_OPTS="--recursive --links --safe-links --perms --times --compress --force --whole-file --delete --stats --timeout=180 --exclude=/distfiles --exclude=/local --exclude=/packages" PORTAGE_TMPDIR="/var/tmp" PORTDIR="/usr/portage" PORTDIR_OVERLAY="/usr/local/portage" SYNC="rsync://rsync.gentoo.org/gentoo-portage" USE="X acl apache2 berkdb cdr cli cracklib crypt cups dri dvd fortran gd gdbm gpm gtk gtk2 iconv ipv6 isdnlog kde kdgraphics midi mudflap ncurses nls nptl nptlonly openmp pam pcre perl pic pppd python qt readline reflection session spl ssl tcpd unicode x86 xorg zlib" ALSA_CARDS="ali5451 als4000 atiixp atiixp-modem bt87x ca0106 cmipci emu10k1 emu10k1x ens1370 ens1371 es1938 es1968 fm801 hda-intel intel8x0 intel8x0m maestro3 trident usb-audio via82xx via82xx-modem ymfpci" ALSA_PCM_PLUGINS="adpcm alaw asym copy dmix dshare dsnoop empty extplug file hooks iec958 ioplug ladspa lfloat linear meter mulaw multi null plug rate route share shm softvol" APACHE2_MODULES="actions alias auth_basic authn_alias authn_anon authn_dbm authn_default authn_file authz_dbm authz_default authz_groupfile authz_host authz_owner authz_user autoindex cache dav dav_fs dav_lock deflate dir disk_cache env expires ext_filter file_cache filter headers include info log_config logio mem_cache mime mime_magic negotiation rewrite setenvif speling status unique_id userdir usertrack vhost_alias" ELIBC="glibc" INPUT_DEVICES="keyboard mouse evdev" KERNEL="linux" LCD_DEVICES="bayrad cfontz cfontz633 glk hd44780 lb216 lcdm001 mtxorb ncurses text" USERLAND="GNU" VIDEO_CARDS="RADEON" Unset: CPPFLAGS, CTARGET, EMERGE_DEFAULT_OPTS, INSTALL_MASK, LANG, LC_ALL, LDFLAGS, LINGUAS, PORTAGE_COMPRESS, PORTAGE_COMPRESS_FLAGS, PORTAGE_RSYNC_EXTRA_OPTS I am pretty new to overlays and ebuilds... so I hope that what is here is okay. Just trying to make things work :)
Created attachment 145385 [details] Updated ebuilds, scripts, and INSTALL.GENTOO The previous ebuild was missing "inherit eutils". This tar also includes updated init/conf files and INSTALL.GENTOO to satisfy bug #204731.
Alright, it compile with this new ebuilds! But there is just a little typo... -BEGIN------------------------ * Messages for package sys-fs/ocfs2-tools-1.3.9: * Read /usr/share/doc/ocfs2-tools-1.3.9/INSTALL.GENTOO.gz for instructions * about how to install, configure and run ocfs2. * GNU info directory index is up-to-date. TORQUE2 sys-fs # less /usr/share/doc/ocfs2-tools-1.3.9/INSTALL.GENTOO.gz /usr/share/doc/ocfs2-tools-1.3.9/INSTALL.GENTOO.gz: No such file or directory ------------------------FIN- The correct file is /usr/share/doc/ocfs2-tools-1.3.9/INSTALL.GENTOO.bz2 so the message should reflect that...
I've just installed and configured ocfs2-tools-1.3.9 and it looks good to me :-) It's simple to setup with drbd having active/active nodes too! Thanks for all the hard work on this one.
When is this going to portage? I'm not meaning to create pressure, I'd just like to test it in a few weeks or so (if it's available then)...
First of all thank you for your hard work. I was able to emerge and use ocfs2-tools-1.3.9 in amd64 (with xen). There is only one small "glitch", the rc script for ocfs2 checks the readiness of several modules required to run ocfs2, but if the modules are built-in in the Kernel this step fails. Maybe you get a good idea on how this could be resolved without having to compile the ocfs/configurefs as module, or comment out the check in the rc.
Created attachment 163408 [details] Updated init script to support ocfs2 when built into kernel or as modules.
Created attachment 163410 [details] Ebuild version bump. Just a renamed copy of 1.3.9 ebuild to support the latest version. Only tested on x86.
Do we patch ocfs2-tools for this heartbeat bug? http://oss.oracle.com/bugzilla/show_bug.cgi?id=944 I'm getting it on two of my systems :-(
(In reply to comment #32) > Do we patch ocfs2-tools for this heartbeat bug? > > http://oss.oracle.com/bugzilla/show_bug.cgi?id=944 > > I'm getting it on two of my systems :-( The patch listed in that bug is a patch to the kernel's fs code. The ocfs2-tools package is just the set of mkfs, fsck, etc. tools. In gentoo the ocfs2 fs code is just contained in the mainline kernel, so I'd suggest that you check for a more updated kernel and/or file the details of your bug with the oracle ocfs2 or gentoo kernel guys.
(In reply to comment #33) > The patch listed in that bug is a patch to the kernel's fs code. Gulp.... so it is! I'd not noticed that :-( Perhaps that's why its not fixed my issue LMAO!
After upgrading to linux-2.6.26-gentoo-r3 from linux-2.6.25-gentoo-r7 Cannot mount ocfs2 volumes with this ebuild any more. It fails with error: Nov 27 20:08:43 falcon-cl3 o2net: accepted connection from node falcon-cl2 (num 2) at 172.16.2.235:7777 Nov 27 20:08:47 falcon-cl3 ocfs2: Cluster stack driver "o2cb" cannot be found Nov 27 20:08:47 falcon-cl3 (16189,0):ocfs2_dlm_init:2636 ERROR: status = -2 Nov 27 20:08:47 falcon-cl3 (16189,0):ocfs2_mount_volume:1208 ERROR: status = -2 Nov 27 20:08:47 falcon-cl3 ocfs2: Unmounting device (8,0) on (node 0) Further investigation show loaded kernel module: ocfs2_stackglue According to: http://lkml.indiana.edu/hypermail/linux/kernel/0803.0/2277.html I should have not problem with classic stack. But it seems that its not true.
(In reply to comment #35) > After upgrading to linux-2.6.26-gentoo-r3 from linux-2.6.25-gentoo-r7 > Cannot mount ocfs2 volumes with this ebuild any more. > It fails with error: > Nov 27 20:08:43 falcon-cl3 o2net: accepted connection from node falcon-cl2 (num > 2) at 172.16.2.235:7777 > Nov 27 20:08:47 falcon-cl3 ocfs2: Cluster stack driver "o2cb" cannot be found > Nov 27 20:08:47 falcon-cl3 (16189,0):ocfs2_dlm_init:2636 ERROR: status = -2 > Nov 27 20:08:47 falcon-cl3 (16189,0):ocfs2_mount_volume:1208 ERROR: status = -2 > Nov 27 20:08:47 falcon-cl3 ocfs2: Unmounting device (8,0) on (node 0) > > Further investigation show loaded kernel module: > ocfs2_stackglue > According to: http://lkml.indiana.edu/hypermail/linux/kernel/0803.0/2277.html I > should have not problem with classic stack. But it seems that its not true. 2.6.26 added preliminary (read not necessarily stable) support for the rhcs (openais/cman/dlm) style cluster communication. I haven't upgraded yet, but if you're still intending to use the ocfs2 cluster stack I believe you'll need to "modprobe ocfs2_stack_o2cb".
Thanks for the ebuild but ... As a total newbie to ocfs2, I had to resolve the following problems to be able to mount an ocfs2 partition : - The /etc/init.d/ocfs2 is only working with ocfs2 compiled as modules - The /etc/init.d/ocfs2 masks the error message by redirecting them to null - The ocfs2console doesn't allow to Configure Nodes from the Cluster menu. - The emerged /etc/ocfs2/cluster.conf is not a running as-is - The /etc/conf.d/ocfs2 contains some values different from original ocfs2 package /etc/init.d/ocfs2: * If all the ocfs2 packages are compiled in the kernel, remove the module check. I finally found the original o2cb had a detailed checking, we should probably maybe reuse. * The result from o2cb_ctl should be displayed ocfs2console * The ocfs2console use o2cb command to load, make online or check the status of the cluster. * As this command o2cb has been replaced by our /etc/init.d/ocfs2 script, the Configure Nodes from the Cluster menu failed with a misleading error message "Could not query the state of th cluster stack. This must be resolved before any OCFS2 filesystemcan be mounted" * I propose to install a reduced and gentooized version of o2cb - without daemon start/stop/restart - in /sbin. The python class o2cb_ctl in ocfs2interface package has to be modified. The O2CB_INIT constant has to point to o2cb. /etc/ocfs2/cluster.conf * The default cluster.conf created by emerge should at least contains the correct configuration of the local node. /etc/conf.d/ocfs2 * The configuration file should contain all the potential variable with a comment. Lot of additional variables seem to exist in the original o2cb. Finally, allowing to boot the root partition on a clusterized partition should be ported from the original ocfs2packaging. It could be fine to have the same integration has LVM and EVMS. Thanks for the already done job Best
Port of o2cb as support wrapped command in /sbin Review of /etc/init.d/ocfs2 and delegation to o2cb Add /etc/init.d/ocfs2 status to get the cluster status
Created attachment 173821 [details] ocfsv2 cluster management This file should be placed in /sbin
Created attachment 173822 [details] rc script to handle ocfsv2 cluster Replace /etc/init.d/ocfs2
Created attachment 173824 [details] ocfs2console modified to use o2cb ocfs2console modified to use o2cb and not /etc/init.d/o2cb This file has to be placed as /usr/lib/python2.5/site-packages/ocfs2interface/o2cb_ctl.py
Created attachment 181426 [details, diff] 1.4.1 ebuild patch + files/ocsf2-limits.patch Compilation problems with latest glibc (?). Unlinked from <limits.h>. Patch attached.
> Compilation problems with latest glibc (?). Unlinked from <limits.h>. Patch > attached. I've not had any issues with the ebuild from Brian Kroth using the following: sys-fs/ocfs2-tools-1.4.1 sys-libs/glibc-2.6.1 Are you using a newer glibc than this one?
(In reply to comment #43) > sys-libs/glibc-2.6.1 > > Are you using a newer glibc than this one? Latest. sys-libs/glibc-2.9_p20081201-r1
Created attachment 181533 [details] Updated init script (checking /proc/config.gz too)
I found problem, but unsure to reproducing it in "stable" environment and in target package to report. I use: sys-apps/portage-2.2_rc23 sys-fs/ocfs2-tools-1.4.1 But same place exists in ocfs2-tools-1.2.1.ebuild: after "keepdir /dlm", portage try to merge this file to current "/dlm" - on working system with mounted ocfs2_dlmfs and fall to error (permission denied). I use next: test -e /dlm || keepdir /dlm - but this may produce packaging problems (IMHO better then merge problem). PS Also ocfs2-tools-1.4.1.tar.gz from Oracle have no "ocfs2cdsl" package included (found in git repository).
(to my last post) test -e "${ROOT}"/dlm || keepdir /dlm
Hi, I've just tried the latest patch and I still get this error: MP -MF ./.client_proto.d -o client_proto.o -c client_proto.c client_proto.c: In function `parse_itemcount': client_proto.c:228: error: `LONG_MAX' undeclared (first use in this function) client_proto.c:228: error: (Each undeclared identifier is reported only once client_proto.c:228: error: for each function it appears in.) client_proto.c:228: error: `UINT_MAX' undeclared (first use in this function) client_proto.c: In function `parse_status': client_proto.c:248: error: `LONG_MIN' undeclared (first use in this function) client_proto.c:248: error: `LONG_MAX' undeclared (first use in this function) client_proto.c:249: error: `INT_MIN' undeclared (first use in this function) client_proto.c:249: error: `INT_MAX' undeclared (first use in this function) make[1]: *** [client_proto.o] Error 1 make[1]: Leaving directory `/var/tmp/portage/sys-fs/ocfs2-tools-1.4.1/work/ocfs2-tools-1.4.1/libo2cb' make: *** [libo2cb] Error 2 This error has only started happening since I upgraded to glibc-2.8_p20080602-r1 It happens on both amd64 and x86 systems running the hardened profile. Rich.
(In reply to comment #48) > I've just tried the latest patch and I still get this error: ok, cancel that... i've done it correctly now and it works now. The fact the "limits" diff has more than one diff inside it confused me :-(
Is this going anywhere, like into the tree or into some overlay?
issue described in #46 , #47 also happens with ebuild 1.3.9. keep /dll fails with permision denied, on amd64 hardened with glibc 2.8_p20080602-r1 and mounted cluster
Created attachment 196229 [details] OCFS2 Tools Ebuild Ebuild version bump. The ocsf2-limits.patch is no longer needed.
I suggest to change 'keepdir /dlm' to 'mkdir "${D}"/dlm' to solve problem, described in Comment #46. This way packaged and installed without problem in current environment on new or running system (just without empty file inside).
Created attachment 196249 [details] ocfs2-tools-1.4.2.ebuild There are fixed Comment #46, Bug #272943, Bug #258122.
Where config taken from ? None in above attachments ... newconfd "${FILESDIR}"/ocfs2.conf ocfs2
(In reply to comment #55) > Where config taken from ? None in above attachments ... > > newconfd "${FILESDIR}"/ocfs2.conf ocfs2 > This file is in the "ocfs2-ebuilds.tar.bz2" tarball listed.
(In reply to comment #56) > (In reply to comment #55) > > Where config taken from ? None in above attachments ... > > > > newconfd "${FILESDIR}"/ocfs2.conf ocfs2 > > > > This file is in the "ocfs2-ebuilds.tar.bz2" tarball listed. > Yes. Or in "raw" overlay. v1.4.3 now. Directly here: http://raw.googlecode.com/svn/trunk/sys-fs/ocfs2-tools/
ocfs2 v1.4.4 is now out... However, Oracle have changed the filename you donwnload to this: ocfs2-1.4.4.tar.gz this means that just changing the ebuild name doesn't help try the new version :-( Also, I can't yet get v1.4.4 to compile :-(
(In reply to comment #58) > ocfs2 v1.4.4 is now out... However, Oracle have changed the filename you > donwnload to this: > > ocfs2-1.4.4.tar.gz > > this means that just changing the ebuild name doesn't help try the new version > :-( > Also, I can't yet get v1.4.4 to compile :-( > That would be the tar for the ocfs2 filesystem itself, which you don't need since it's in the mainline kernel. If you want to update that piece, update your kernel. These ebuilds are for the ocfs2-tools package (mkfs.ocfs2(8), fsck.ocfs2(8), tunefs.ocfs2(8), mount.ocfs2(8) and debugfs.ocfs2(8), etc.). The latest version for that is 1.4.3. http://oss.oracle.com/projects/ocfs2-tools/files/source/v1.4/ I haven't tested it but you should be able to rename the current ebuild to 1.4.3 and try to build it. If it doesn't work, post the errors. Brian
Removed from Portage
(In reply to comment #60) > Removed from Portage > Why have you removed this from portage? version 1.4.2 works just fine, someone just needs to include it into the tree and it will solve this whole issue. Infact, the 1.4.2 ebuild on this bug report works also for 1.4.3! Just because you don't understand how to fix it, doesn't mean its broken!
(In reply to comment #61) > > Removed from Portage ... > Just because you don't understand how to fix it, doesn't mean its broken! IMHO maintainer problem... I "supporting" this ebuild for own usage and publish in overlay, but not yet registered as developer and not teach about maintainer/developer realm. And not ready (IMHO) to be always in-touch on public to start learn this process...
> IMHO maintainer problem... > I "supporting" this ebuild for own usage and publish in overlay, but not yet > registered as developer and not teach about maintainer/developer realm. And not > ready (IMHO) to be always in-touch on public to start learn this process... > I would suggest leaving the package there rather than removing it without notice. There are other packages that have no maintainer and they are still in the tree. Why not leave it there until someone claims it, especially as some of us are still using it on a daily basis! Thank you for your help, but please put this back in the tree. Rich
