Submitted by ghomem on Sun, 11/30/2014 - 06:40
After the creation of a Centos 7 guest from the Cloudmin x64 machine template we get:
1- no hd expansion
Expanding filesystem to 40 GB .................................................................................. ................................................................................ ................................................................................ ................................................................................ ...................................................................... .. expansion failed : e2fsck -f -p \/dev\/loop1 failed : /dev/loop1 has unsupported feature(s): 64bit e2fsck: Get a newer version of e2fsck!
2- no IP address set on eth0
Status:
Closed (fixed)
Comments
Submitted by ghomem on Sun, 11/30/2014 - 06:47 Comment #1
Hmmm, point nr 2) is probably due to a new interface naming style on RHEL / Centos 7. If so, please disregard.
Submitted by JamieCameron on Sun, 11/30/2014 - 11:53 Comment #2
Point 1 happens because the centos 6 host doesn't support the 64 bit ext4 filesystem on the VM. Unfortunately there isn't any easy solution other than upgrading the host, or its ext tools package.
Submitted by ghomem on Sun, 11/30/2014 - 12:05 Comment #3
Well, "upgrading the host", that possibly hosts tens of VMs, isn't something that can be suggested lightly, is it? I hope you are not saying that it won't be possible to run Centos 7 in Centos 6 based hosts.
I don't want to believe that, but even if true Centos 7 is not supported as a host. From install.sh:
Currently supported systems:prosupported=" CentOS and RHEL 5-6 on i386 and x86_64 Scientific Linux 6 on i386 and x86_64 Debian 6.0 and 7.0 on i386 and amd64 Ubuntu 8.04 LTS, 10.04 LTS, and 12.04 LTS on i386 and amd64 Amazon Linux 2012.03 on i386 and x86_64 FreeBSD 7 and 8 on i386 and amd64"
Regarding the upgrade of ext-tools: if that is something you can provide and it's proven to work please let us know.
Submitted by JamieCameron on Sun, 11/30/2014 - 12:27 Comment #4
Right, upgrading an existing system is pretty risky - what I meant was to start with CentOS 7 in the first place or migrate VMs to a new host.
Another alternative is to create a CentOS 7 image that uses ext3 as the filesystem. Are you using Cloudmin-provided images there, or did you create your own by doing an install into an empty VM?
Submitted by ghomem on Sun, 11/30/2014 - 12:44 Comment #5
As stated above:
"Centos 7 guest from the Cloudmin x64 machine template"
I tried to use your template as we do with Centos 6. One advantage of Cloudmin is providing these minimal templates so that is what I used.
I'm not too much in a hurry regarding this but there will be a time when only Centos 7 makes sense for new VMs and upgrading the hosts is not reasonable.
Note:
If you are going to look at the template perhaps you should consider having ii using the default interface naming (ethX) and including the net-tools package (ifconfig). I say consider because I don't want to be too assertive on this since I'm just starting too look at Centos 7 now. I wonder what good reasons there are to break things that work like these.
Submitted by JamieCameron on Mon, 12/01/2014 - 17:36 Comment #6
I think what we'll do is create a new CentOS 7 template that uses the ext3 filesystem (and ethX style interface names) to avoid this problem completely.
Submitted by ghomem on Mon, 12/01/2014 - 17:45 Comment #7
Sounds fine with me, we don't run anything so tight on performance that a milisecond of difference due to a better filesystem would be relevant, if that's even the case.
Submitted by JamieCameron on Mon, 12/01/2014 - 22:41 Comment #8
Ok, an updated version (1.2) of the CentOS 7 image is now available, which uses an EXT3 filesystem instead of EXT4.
Submitted by ghomem on Sat, 02/04/2017 - 18:04 Comment #9
Submitted by IssueBot on Thu, 10/11/2018 - 20:07 Comment #10
Automatically closed - issue fixed for 2 weeks with no activity.