Bug 1984740
| Summary: | Use of inst.selinux=0 fails to set root password | ||
|---|---|---|---|
| Product: | Red Hat Enterprise Linux 8 | Reporter: | Parikshit Khedekar <pkhedeka> |
| Component: | shadow-utils | Assignee: | Iker Pedrosa <ipedrosa> |
| Status: | CLOSED ERRATA | QA Contact: | Anuj Borah <aborah> |
| Severity: | medium | Docs Contact: | |
| Priority: | high | ||
| Version: | 8.4 | CC: | aboscatt, derli, dwalsh, ipedrosa, jkonecny, jstodola, lopresti, lvrabec, mmalik, paygupta, pbrezina, plautrba, tscherf, vmojzis |
| Target Milestone: | beta | Keywords: | Triaged |
| Target Release: | --- | ||
| Hardware: | All | ||
| OS: | Linux | ||
| Whiteboard: | |||
| Fixed In Version: | shadow-utils-4.6-18.el8 | Doc Type: | If docs needed, set a value |
| Doc Text: | Story Points: | --- | |
| Clone Of: | Environment: | ||
| Last Closed: | 2023-11-14 15:49:14 UTC | Type: | Bug |
| Regression: | --- | Mount Type: | --- |
| Documentation: | --- | CRM: | |
| Verified Versions: | Category: | --- | |
| oVirt Team: | --- | RHEL 7.3 requirements from Atomic Host: | |
| Cloudforms Team: | --- | Target Upstream Version: | |
| Embargoed: | |||
I was looking into this and attaching log. The interesting part: 10:07:49,944 NOTICE kernel:audit: type=1107 audit(1663754869.943:1525): pid=53923 uid=0 auid=4294967295 ses=4294967295 subj=system_u:system_r:kernel_t:s0 msg='avc: could not determine enforcing mode: No such file or directory#012 exe="/usr/sbin/chpasswd" sauid=0 hostname=? addr=? terminal=?' 10:07:49,944 NOTICE kernel:audit: type=1701 audit(1663754869.943:1526): auid=4294967295 uid=0 gid=0 ses=4294967295 subj=system_u:system_r:kernel_t:s0 pid=53923 comm="chpasswd" exe="/usr/sbin/chpasswd" sig=6 res=1 10:07:49,983 INFO systemd:Created slice system-systemd\x2dcoredump.slice. 10:07:49,986 INFO systemd:Started Process Core Dump (PID 53924/UID 0). 10:07:49,987 NOTICE kernel:audit: type=1130 audit(1663754869.985:1527): pid=1 uid=0 auid=4294967295 ses=4294967295 subj=system_u:system_r:kernel_t:s0 msg='unit=systemd-coredump@0-53924-0 comm="systemd" exe="/usr/lib/systemd/systemd" hostname=? addr=? terminal=? res=success' 10:07:50,169 INFO systemd-coredump:Resource limits disable core dumping for process 53923 (chpasswd). 10:07:50,176 CRIT systemd-coredump:Process 53923 (chpasswd) of user 0 dumped core. Seems like the chpasswd call will crash which could be potential SELinux issue even though the SELinux should be set to permissive. Based on the first line it seems that it has an issue to determine the enforcing mode. Based on the above I would like to know SELinux team opinion on why avc can't determine enforcing mode and if that could be the root cause of the crash. Could this be an issue that chpasswd is started with chroot -R option for chroot where the SELinux is not completely set. However, installation environment is set to Permissive. chpasswd call: chpasswd -R /mnt/sysroot -e $ getenforce Permissive Switching to SELinux team for further investigation. It looks like a problem similar to https://bugzilla.redhat.com/show_bug.cgi?id=1993081 which seems to be already fixed in shadow-utils rhel9. The problem with `chpasswd -R ...` is that libselinux is initialized before `chroot()` as enabled. But when the process `chroot()`s there's no /sys/fs/selinux in the chroot and libselinux is confused. IIRC the solution was to call `selabel_close()` in order to reset libselinux state, or something. CC @ipedrosa The (In reply to Petr Lautrbach from comment #14) > It looks like a problem similar to > https://bugzilla.redhat.com/show_bug.cgi?id=1993081 which seems to be > already fixed in shadow-utils rhel9. The selinux labeling code in shadow-utils changed before RHEL9 and that's why the regression happened. You can take a look at the code before: https://github.com/shadow-maint/shadow/blob/e6416fd81bff205f9daac3da02cb3392d3d75e58/lib/selinux.c#L54 And after those changes: https://github.com/shadow-maint/shadow/blob/master/lib/selinux.c#L47 I don't think those changes are affecting RHEL8 because I didn't port them. > The problem with `chpasswd -R ...` is that libselinux is initialized before > `chroot()` as enabled. But when the process `chroot()`s there's no > /sys/fs/selinux in the chroot and libselinux is confused. IIRC the solution > was to call `selabel_close()` in order to reset libselinux state, or > something. > > CC @ipedrosa I don't see any question and I'm not an expert in selinux. So, how can I help you? It wasn't a question. I used a wrong CC technology to let you know about this issue releated to chpasswd. Setup: # dnf --instalroot=/tmp/root group install 'Minimal Install' # mount -t proc proc /tmp/root/proc # setenforce 0 RHEL8 version: [root@rhel8 ~]# echo 'root:e' | chpasswd -R /tmp/root chpasswd: avc.c:74: avc_context_to_sid_raw: Assertion `avc_running' failed. Aborted (core dumped) RHEL9 version: [root@rhel9 ~]# echo 'root:e' | chpasswd -R /tmp/root [root@rhel9 ~]# Since the problem described in this bug report should be resolved in a recent advisory, it has been closed with a resolution of ERRATA. For information on the advisory (Low: shadow-utils security and bug fix update), and where to find the updated files, follow the link below. If the solution does not work for you, open a new bug report. https://access.redhat.com/errata/RHSA-2023:7112 |
Description of problem: When we use "inst.selinux=0" to install any 8.X it fails with the traceback as, ------------ ... Configuring installed system ............... An unknown error has occured, look at the /tmp/anaconda-tb* file(s) for more details sync_run_task(self._task_proxy) File "/usr/lib64/python3.6/site-packages/pyanaconda/installation_tasks.py", line 490, in start self.run_task() File "/usr/lib64/python3.6/site-packages/pyanaconda/installation_tasks.py", line 311, in start item.start() File "/usr/lib64/python3.6/site-packages/pyanaconda/installation_tasks.py", line 311, in start item.start() File "/usr/lib64/python3.6/site-packages/pyanaconda/installation_tasks.py", line 311, in start item.start() File "/usr/lib64/python3.6/site-packages/pyanaconda/installation.py", line 403, in run_installation queue.start() File "/usr/lib64/python3.6/threading.py", line 864, in run self._target(*self._args, **self._kwargs) File "/usr/lib64/python3.6/site-packages/pyanaconda/threading.py", line 280, in run threading.Thread.run(self) dasbus.error.DBusError: Unable to set password for new user: status=-6 ------------ Checking it deep I see the command fails as, [anaconda root@localhost ~]# chpasswd -R /mnt/sysimage/ -e chpasswd: avc.c:74: avc_context_to_sid_raw: Assertion `avc_running' failed. Aborted (core dumped) [anaconda root@localhost ~]# Gathered the dumps as well, but not able to make solid conclusion from it as I'm not that expert to selinux yet but I fee the patch shadow-4.3.1-selinux-perms.patch introduces the changes in RHEL 8 whereas it works fine on RHEL7 Version-Release number of selected component (if applicable): RHEL 8.X How reproducible: Use inst.selinux=0 to boot line as, BOOT_IMAGE=images/8.1/vmlinuz initrd=images/8.1/initrd.img inst.stage2=hd:LABEL=RHEL-8-1-0-BaseOS-x86_64 inst.ks=ftp://192.168.122.1/pub/rhel8.4.ks inst.selinux=0 Actual results: Fails to set a root password and results to traceback Expected results: Installation shall work fine. Additional info: The current workaround is to set the selinux=0 as well, I don't see inst.selinux=0 turning off selinux as it still marks it permissive. There were confusion in RH doc and Upstream doc so based on theory I have asked for a change to upstream doc as, https://github.com/rhinstaller/anaconda/pull/3530 I'm not sure if this needs shadow-utils as well, but as this is install time I'm marking it to anaconda