#348 Issue closed: Rear dependencies on RHEL7 beta 1

Labels: support / question

Florent38 opened issue at 2013-12-17 08:00:

Hi

I'm trying to install rear on RHEL 7 beta 1 with the latest git version, but the mingetty package no longer exist on the new RHEL... But agetty is present.
Have you any idea to get around this problem ?

Thanks

gdha commented at 2013-12-17 15:02:

is mingetty not available in yum repo's?

Florent38 commented at 2013-12-18 07:59:

Mingetty is not available in yum repo's...
I downloaded mingetty of Fedora 19 to try to make a backup / restore.

gdha commented at 2013-12-18 08:11:

Perhaps open a bugzilla report against RHEL 7.0 beta1 for the missing mingetty? I'm sure we are not the only project depending on it...

Florent38 commented at 2013-12-18 12:16:

According to RHEL is in beta version, I'll wait a bit before opening a bugzilla.

Florent38 commented at 2013-12-18 13:53:

Rear restored RHEL7 beta1 succesfully. I just added the f option to mkfs to force recreation of xfs filesystems (See #349)

gdha commented at 2013-12-18 14:27:

submitted a bugzilla request myself - https://bugzilla.redhat.com/show_bug.cgi?id=1044546

gdha commented at 2013-12-20 09:02:

@Florent38 did a test on RHEL7 myself - still have some remarks:

  1. mkfs -f cannot work as mkfs does not know about the parameter -f. It refers to mkfs.xfs I guess. Therefore, the solution mentioned in issue #349 is not working for me.
  2. RHEL7 can work without mingetty and can use the built-in agetty without problems. To accomplish this I needed to remove mingetty from the list of REQUIRED_PROGS (is ok as it is also listed in PROGS)
    However, we can still enforce mingetty RPM via the rear.spec file (except for RHEL7 then). I will close the bugzilla report as we have a valid work-around.
  3. a warning that needs some attention:
2013-12-19 15:49:41 Start system layout restoration.
  /run/lvm/lvmetad.socket: connect failed: No such file or directory
  WARNING: Failed to connect to lvmetad: No such file or directory. Falling back to internal scanning.

jhoekx commented at 2013-12-20 09:06:

About 3. Adding support for it wouldn't be hard, but it's not strictly needed. It's a daemon that manages LVM metadata, so LVM does not have to scan for that metadata every time there is an LVM operation. It just improves performance.

gdha commented at 2013-12-20 14:56:

Just pull in a view changes (forgot to link to this issue) for pts 1, 2 and 3.


[Export of Github issue for rear/rear.]