| Summary: | sys-cluster/cluster-glue-1.0.12-r1 stable request (was: sys-cluster/cluster-glue fails with: useradd: invalid shell '/dev/null') | ||||||
|---|---|---|---|---|---|---|---|
| Product: | Gentoo Linux | Reporter: | Sebastian Pipping <sping> | ||||
| Component: | Stabilization | Assignee: | Gentoo Cluster Team <cluster> | ||||
| Status: | RESOLVED FIXED | ||||||
| Severity: | normal | CC: | bugs, earny, merlin, sandino, seen_xu, treecleaner | ||||
| Priority: | Normal | Keywords: | CC-ARCHES, STABLEREQ | ||||
| Version: | unspecified | Flags: | nattka:
sanity-check+
|
||||
| Hardware: | All | ||||||
| OS: | Linux | ||||||
| Whiteboard: | |||||||
| Package list: |
sys-cluster/cluster-glue-1.0.12-r1
|
Runtime testing required: | --- | ||||
| Bug Depends on: | |||||||
| Bug Blocks: | 711674, 747415 | ||||||
| Deadline: | 2020-12-07 | ||||||
| Attachments: |
|
||||||
Same with 1.0.11 and 1.0.12. useradd is from sys-apps/shadow-4.8-r2 . Same with sys-apps/shadow-4.8-r3::gentoo same here sys-apps/shadow 4.8-r4 sys-cluster/cluster-glue 1.0.12 Adding shell /dev/null to /etc/shells is ignored. useradd does not read this file (anymore?). useradd tests /bin/null for exec rights and fails (of course). Adding user "hacluster" manually with "/bin/false" as shell mitigates this problem. The bug has been referenced in the following commit(s): https://gitweb.gentoo.org/repo/gentoo.git/commit/?id=f51b83a43a70a06d93851b0fa41f7e6e993e1e6e commit f51b83a43a70a06d93851b0fa41f7e6e993e1e6e Author: Michał Górny <mgorny@gentoo.org> AuthorDate: 2020-11-07 09:21:38 +0000 Commit: Michał Górny <mgorny@gentoo.org> CommitDate: 2020-11-07 09:21:48 +0000 package.mask: Last rite sys-cluster/cluster-glue & revdeps Bug: https://bugs.gentoo.org/704610 Signed-off-by: Michał Górny <mgorny@gentoo.org> profiles/base/package.use.mask | 4 ++++ profiles/package.mask | 14 ++++++++++++++ 2 files changed, 18 insertions(+) i got this to build just fine. why is this slated for removal?? no problem with: sys-cluster/cluster-glue-1.0.12-r1 sys-cluster/resource-agents-3.9.6 sys-cluster/pacemaker-2.0.4 also my test machine don't have python-2. and stuff is still ok. pacemaker-2.0.5 is in RC2 state (soon to be released officially) Like Zdravko Spoljar , we also have no problem, emerging and get running pacemaker emerge pacemaker corosync cluster-glue crmsh resource-agents -pv These are the packages that would be merged, in order: Calculating dependencies... done! [ebuild R ~] sys-cluster/cluster-glue-1.0.12-r1::gentoo USE="-doc -ipmilan -libnet -static-libs" 0 KiB [ebuild R ~] sys-cluster/corosync-3.0.4::gentoo USE="xml -dbus -doc -systemd" 0 KiB [ebuild R ] sys-cluster/resource-agents-3.9.6::gentoo USE="-doc -libnet -rgmanager" 0 KiB [ebuild R ~] sys-cluster/pacemaker-2.0.4::gentoo USE="acl -smtp -snmp" PYTHON_SINGLE_TARGET="python3_6 -python3_7 -python3_8 -python3_9" 0 KiB [ebuild R ~] sys-cluster/crmsh-4.1.0::gentoo PYTHON_TARGETS="python3_6 python3_7" 0 KiB Total: 5 packages (5 reinstalls), Size of downloads: 0 KiB Due to missing alternatives we strongly recommend not removing pacemaker to provide the ability to set up HA clusters! Yeah, removing pacemaker from gentoo would be a problem The thing is mostly the cluster team is not so active I for once don't use / follow pacemaker any more so I'm not a good citizen in making sure it's clean etc I guess that's the reason why we got to this point, I've merged a github PR to try to help but this got me 5 open QA bugs... Pacemaker et al got their last rites notice and to my understanding nobody claimed them. If you want to save it please add yourself as maintainer and remove the cluster team. i'm not quite sure that i have sufficient knowledge about procedures&ebuilds to be maintainer. but for first shot, how about removing just 1.x and keep 2.x under ~amd64 ? and where to look for package list which "belong" to "cluster team" ? (In reply to Zdravko Spoljar from comment #9) > and where to look for package list which "belong" to "cluster team" ? # find /usr/portage/ -mindepth 3 -maxdepth 3 -type f -name metadata.xml -exec fgrep -l cluster@gentoo.org {} + 2>/dev/null | sort The bug has been referenced in the following commit(s): https://gitweb.gentoo.org/repo/gentoo.git/commit/?id=0275c17207295dced2f8f1d68f357e443a8f2aaa commit 0275c17207295dced2f8f1d68f357e443a8f2aaa Author: Sebastian Pipping <sping@gentoo.org> AuthorDate: 2020-11-12 11:06:28 +0000 Commit: Sebastian Pipping <sping@gentoo.org> CommitDate: 2020-11-12 11:19:48 +0000 package.mask: Revert last rite of sys-cluster/cluster-glue & revdeps As discussed with Marc Schiffbauer <mschiff@gentoo.org>. This reverts commit f51b83a43a70a06d93851b0fa41f7e6e993e1e6e. Bug #704610 of sys-cluster/cluster-glue turned out to be fixed by commit b5442dd701a9eaaf22fb92808fb0ec93f7a9f1e6 of July 2020. Vulnerable sys-cluster/pacemaker has been bumped yesterday. So the path to stabilization is no longer blocked. Bug: https://bugs.gentoo.org/704610 Bug: https://bugs.gentoo.org/711674 Bug: https://bugs.gentoo.org/751430 Signed-off-by: Sebastian Pipping <sping@gentoo.org> profiles/base/package.use.mask | 4 ---- profiles/package.mask | 14 -------------- 2 files changed, 18 deletions(-) x86 stable hppa stable amd64 done all arches done |
Created attachment 602348 [details] Build log Installation fails, see build log. [..] * Adding user 'hacluster' to your system ... * - Userid: 999 * - Shell: /dev/null * - Home: /var/lib/heartbeat * - Groups: haclient * - GECOS: added by portage for cluster-glue useradd: invalid shell '/dev/null' [..]