Summary: | app-emulation/libvirt-5.8.0 fails to start domain with "Unable to get host boot time: No such process" | ||
---|---|---|---|
Product: | Gentoo Linux | Reporter: | ron widler <ron> |
Component: | Current packages | Assignee: | Matthias Maier <tamiko> |
Status: | RESOLVED FIXED | ||
Severity: | normal | CC: | gentoo-bugs, mike, o.freyermuth, ron, virtualization |
Priority: | Normal | Keywords: | PATCH |
Version: | unspecified | ||
Hardware: | All | ||
OS: | Linux | ||
Whiteboard: | |||
Package list: | Runtime testing required: | --- | |
Bug Depends on: | 707982 | ||
Bug Blocks: | |||
Attachments: |
patches virhostuptime.c for host boot time fallback
updated ebuild for libvirt-5.8.0 emerge info emerge --info of correct emerge --info of failed strace of correct strace of failed |
Description
ron widler
2019-10-23 23:55:26 UTC
Created attachment 593730 [details]
updated ebuild for libvirt-5.8.0
Created attachment 593732 [details]
emerge info
I have same problem with one of my gentoo systems but attached patch didn`t help. Attached some logs from "correct" and "failed" systems. At first sight main difference between systems is mdadm+dracut on "failed" and none of them on "correct". Created attachment 594220 [details]
emerge --info of correct
Created attachment 594222 [details]
emerge --info of failed
Created attachment 594224 [details]
strace of correct
Created attachment 594226 [details]
strace of failed
Also made minimal working example, that simulates boot time:
#include <utmpx.h>
#include <errno.h>
#include <stddef.h>
#include <stdio.h>
static unsigned long long bootTime;
static int bootTimeErrno;
void
main(void)
{
struct utmpx id = {.ut_type = BOOT_TIME};
struct utmpx *res = NULL;
if (!(res = getutxid(&id))) {
bootTimeErrno = errno;
} else {
bootTime = res->ut_tv.tv_sec;
}
printf("bootTime %lld\n", bootTime);
printf("bootTimeErrno %d\n", bootTimeErrno);
endutxent();
}
I can confirm this one. the suggestion from https://bugzilla.redhat.com/show_bug.cgi?id=1760885#c1 helped me: disabling remember_owner in /etc/libvirt/qemu.conf and restarting libvirtd The bug has been closed via the following commit(s): https://gitweb.gentoo.org/repo/gentoo.git/commit/?id=bb4fc1071be2c3cc4db7f3ce9ba4b5ef15aae3ac commit bb4fc1071be2c3cc4db7f3ce9ba4b5ef15aae3ac Author: Matthias Maier <tamiko@gentoo.org> AuthorDate: 2019-12-31 21:07:11 +0000 Commit: Matthias Maier <tamiko@gentoo.org> CommitDate: 2019-12-31 22:01:21 +0000 app-emulation/libvirt: version bump to 5.10.0 Closes: https://bugs.gentoo.org/698424 Closes: https://bugs.gentoo.org/702816 Closes: https://github.com/gentoo/gentoo/pull/14190 Package-Manager: Portage-2.3.83, Repoman-2.3.20 Signed-off-by: Matthias Maier <tamiko@gentoo.org> app-emulation/libvirt/Manifest | 1 + app-emulation/libvirt/libvirt-5.10.0.ebuild | 379 ++++++++++++++++++++++++++++ 2 files changed, 380 insertions(+) STILL A BUG. app-emulation/libvirt-5.8.0 just appeared in my "stable" listing, so I emerge'd it. Now, none of my KVM/QEMU guests will start. So, for "stable" users, this bug still exists. (In reply to Gordon Pritchard from comment #10) > STILL A BUG. ***__UNDERSTOOD__*** Upgrade to 5.8.0 also broke mine. Bumped to 5.10.0 and immediate resolution. Suggest 5.8.0 gets blacklisted somehow 'till 5.10.0 is made stable. |