Hi,
When creating new set of rpms for centos-6.3, I just found a problem when imaging clients in the initrd image.
In fact, it looks like I'm hitting a bug very similar to:
https://bugzilla.redhat.com/show_bug.cgi?id=441244
This can be overrided by replacing line 578 in Server.pm:
$startMB{$m} = q($END_OF_LAST_LOGICAL);
by
$startMB{$m} = q#$(echo "scale=3; ($END_OF_LAST_LOGICAL + 1)" | bc)#;
But this is a ugly patch as you loose 1 block on the disk.
Now I think that the latest parted (v3.1) may fix this issue.
anyway, fixed or not, I4ve successfully bult a new set of rpms with v3.1.
Here is the small patch the build v3.1 instead of v2.3 (attached).
Also attached, the ugly patch (which IMHO, shouldn't be pushed as IMHO, parted v3.1 fixes it).
If someone with write access to the SVN think it's worth pushing the parted-v3.1 patch, please do so.
I'll test ASAP and report.
Cheers,
PS: With this RPMS I was able to fully deploy a beta OSCAR cluster on centos-6.3 (hardware: tyan motherboards with 7 M2090 GPUs and HP DL165 G7
PPS: I'm using post install scripts to replace missing components of systeconfigurator (network config, grub install, initramfs update)
--
Olivier LAHAYE
CEA DRT/LIST/DCSI/DIR
|
