Hi, Observing this since May 2019 in different machines. 1. A fresh installation of Fedora 29 or Fedora 30. 2. Clone ceph repo. 3. ./install-deps.sh 4. Reboot 5. The login screen go in loops. In some instances, the booting never reach the login screen. I remember I was able to boot to single user mode the first time I saw this, but my attempts didn't fix the problem back then with the GUI. It doesn't happen with CentOS Linux release 7.7.1908 (Core). So I use CentOS nowadays. Regards, Jos Collin _______________________________________________ Dev mailing list -- dev(a)ceph.io To unsubscribe send an email to dev-leave(a)ceph.io

This happened on both my laptop and dev box a while back... I didn't know it was related to something that install-deps.sh was doing though. s On Sat, 5 Oct 2019, Brad Hubbard wrote: > BTW, I should probably mention that I have not seen this issue with > ceph before but I have seen it on several occasions. I just assumed it > was a bad system update. I guess in your case it could be triggered by > the install-deps.sh script but, if it is the selinux issue, it's not > really related to ceph directly AFAIUI. > > On Sat, Oct 5, 2019 at 4:49 PM Brad Hubbard <bhubbard(a)redhat.com> wrote: > > > > Try "selinux=0" at the end of the kernel command line. If that works > > create a file called ".autorelabel" in the root of each filesystem > > (usually just '/') and reboot. > > > > On Sat, Oct 5, 2019 at 2:03 PM Jos Collin <jcollin(a)redhat.com> wrote: > > > > > > Hi, > > > > > > Observing this since May 2019 in different machines. > > > > > > 1. A fresh installation of Fedora 29 or Fedora 30. > > > 2. Clone ceph repo. > > > 3. ./install-deps.sh > > > 4. Reboot > > > 5. The login screen go in loops. > > > > > > In some instances, the booting never reach the login screen. I remember > > > I was able to boot to single user mode the first time I saw this, but my > > > attempts didn't fix the problem back then with the GUI. > > > > > > It doesn't happen with CentOS Linux release 7.7.1908 (Core). So I use > > > CentOS nowadays. > > > > > > Regards, > > > Jos Collin > > > _______________________________________________ > > > Dev mailing list -- dev(a)ceph.io > > > To unsubscribe send an email to dev-leave(a)ceph.io > > > > > > > > -- > > Cheers, > > Brad > > > > -- > Cheers, > Brad > _______________________________________________ > Dev mailing list -- dev(a)ceph.io > To unsubscribe send an email to dev-leave(a)ceph.io > >

On Sun, Oct 6, 2019 at 10:19 AM Brad Hubbard <bhubbard(a)redhat.com> wrote: probably we could rule out the suspicion on install-deps.sh by logging on using another account?

Working out what the selinux denial error is should help to shed light. The fact that relabeling the filesystem resolves the issue could indicate that something in the past labelled it incorrectly to begin with. I don't doubt running install-deps.sh could trigger this issue but I strongly suspect the root cause lies with some bad system package or update that results in an incorrectly labelled filesystem. Just my 2c. In order to 'know' we would have to understand the problem a lot better.