Bug 1298234

Summary:	systemd-nspawn container fails to boot for the first time when --link-journal=try-guest is used
Product:	Red Hat Enterprise Linux 7	Reporter:	Frantisek Sumsal <fsumsal>
Component:	systemd	Assignee:	systemd-maint
Status:	CLOSED WONTFIX	QA Contact:	qe-baseos-daemons
Severity:	unspecified	Docs Contact:
Priority:	unspecified
Version:	7.2	CC:	mscherer, msekleta, systemd-maint-list
Target Milestone:	rc
Target Release:	---
Hardware:	Unspecified
OS:	Unspecified
Whiteboard:
Fixed In Version:		Doc Type:	If docs needed, set a value
Doc Text:		Story Points:	---
Clone Of:		Environment:
Last Closed:	2020-12-15 07:39:34 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:

Description Frantisek Sumsal 2016-01-13 15:04:38 UTC

Description of problem:
First nspawn container boot fails when --link-journal=try-guest is specified as an argument of systemd-nspawn (which is used in systemd-nspawn@.service file).

Version-Release number of selected component (if applicable):
systemd-219-19.el7.x86_64

How reproducible:
always

Steps to Reproduce:
1. # mkdir -p /var/lib/machines/testvm
2. # yum -y --installroot="/var/lib/machines/testvm" groupinstall "Minimal install"
3. # systemd-nspawn --link-journal=try-guest -bM testvm

Actual results:
# systemd-nspawn --link-journal=try-guest -bM testvm
Spawning container testvm on /var/lib/machines/testvm.
Press ^] three times within 1s to kill container.
Container testvm failed with error code 1.

Expected results:
Container should boot successfully.

Additional info:
Same process works on Fedora 23:

# mkdir -p /var/lib/machines/testvm
# dnf -y --installroot="/var/lib/machines/testvm" --releasever 23 groupinstall "Minimal install"
<...>
# systemd-nspawn --link-journal=try-guest -bM testvm
Spawning container testvm on /var/lib/machines/testvm.
Press ^] three times within 1s to kill container.
systemd 222 running in system mode. (+PAM +AUDIT +SELINUX +IMA -APPARMOR +SMACK +SYSVINIT
 +UTMP +LIBCRYPTSETUP +GCRYPT +GNUTLS +ACL +XZ -LZ4 +SECCOMP +BLKID +ELFUTILS +KMOD +IDN)
Detected virtualization systemd-nspawn.
Detected architecture x86-64.

Welcome to Fedora 23 (Twenty Three)!
<...>

Comment 2 M. Scherer 2016-04-21 20:46:46 UTC

So I can reproduce this one, and I can also say this is fixed in Fedora 23.

I suspect this work on the first time due to 8054d749c4ad69503b5b2735864f8e72a1b73e62

Comment 3 M. Scherer 2016-04-21 23:10:01 UTC

So I did took a closer look, because it did bother me a lot, but gdb seems to interfere with systemd nspawn, as my error point to "open_console", and I think that's because i run in gdb. I tried with gdbserver, but also didn't got far.

Comment 4 M. Scherer 2016-04-22 09:51:20 UTC

So the problem likely come from /etc/machine-id failing to be created in the container. 

After using strace properly, i see that it fail with "failed to parse machine-id".


This was likely fixed by e01ff70a77e78.

Comment 5 M. Scherer 2016-04-22 10:29:13 UTC

Ok so the fix/workaround is to run  systemd-machine-id-setup --root, or systemd-firstboot --root.

Comment 6 Michal Sekletar 2016-04-22 12:43:07 UTC

We call systemd-machine-id-setup from systemd %post during rpm installation. Hence /etc/machine-id should be present in the container after yum has finished installing packages.

Comment 7 M. Scherer 2016-04-22 13:04:00 UTC

yeah, but it is silently ignoring all errors. 

systemd-machine-id-setup >/dev/null 2>&1 || :


I just tried again on a centos 7, and the machine-id file is empty.

Comment 8 Michal Sekletar 2016-04-22 14:25:46 UTC

I've backported 8054d749c4ad69503b5b2735864f8e72a1b73e62. Feel free to grab test packages.

http://people.redhat.com/~msekleta/bz1298234/

Comment 11 RHEL Program Management 2020-12-15 07:39:34 UTC

After evaluating this issue, there are no plans to address it further or fix it in an upcoming release.  Therefore, it is being closed.  If plans change such that this issue will be fixed in an upcoming release, then the bug can be reopened.