Using latest e2fsprogs on various offline VM ext filesystems

Using latest e2fsprogs on various offline VM ext filesystems

[Daniel Johnson]

Apologies if this "user" question has been asked before.

Tomorrow I will be stopping a couple of dozen kvm virtual machines
running various Linux distros and running a forced e2fsck on all of
their ext filesystems. The distros are a mix of various flavours of
Ubuntu, Debian and Centos, so will include some fs created with older
kernels and E2fsprogs.

The host is 64-bit Centos 6 running vanilla kernel 3.14.39. I figured
I'll build the latest E2fsprogs 1.42.12 for this rather than using the
Centos 6 e2fsprogs-1.41.12-21.el6.x86_64.

My question is; should it be safe in theory to use the latest
E2fsprogs on a mix of ext filesystems in this way? I would have
thought so but don't want to rely on any assumptions :-)

I'll be setting up device maps on the host server to get to the
filesystems on each VMs virtual disk, and doing read only checks. But
if anything needs a repair then I could potentially have a fs modified
on newer kernel & userspace that then has to boot inside a VM with a
distro with much older kernel and userspace such as Centos 5, which is
what I'm concerned about.

Thanks,
Dan

Re: Using latest e2fsprogs on various offline VM ext filesystems

[Andreas Dilger]

You can safely use newer e2fsprogs on older file systems. This should not
affect compatibility with older kernels in any way, unless you explicitly
enable features.

Note that if you are using LVM/DM you could create a snapshot of the
running device and run e2fsck on that. If no errors are found then you
don't need to stop the VM at all. You can reset the last checked counter
and mount count via tune2fs. Only if you find an error in the snapshot do
you need to unmount the filesystem to fix it.

Cheers, Andreas

> On Apr 18, 2015, at 06:46, Daniel Johnson <dan.u.johns@gmail.com> wrote:
> 
> Apologies if this "user" question has been asked before.
> 
> Tomorrow I will be stopping a couple of dozen kvm virtual machines
> running various Linux distros and running a forced e2fsck on all of
> their ext filesystems. The distros are a mix of various flavours of
> Ubuntu, Debian and Centos, so will include some fs created with older
> kernels and E2fsprogs.
> 
> The host is 64-bit Centos 6 running vanilla kernel 3.14.39. I figured
> I'll build the latest E2fsprogs 1.42.12 for this rather than using the
> Centos 6 e2fsprogs-1.41.12-21.el6.x86_64.
> 
> My question is; should it be safe in theory to use the latest
> E2fsprogs on a mix of ext filesystems in this way? I would have
> thought so but don't want to rely on any assumptions :-)
> 
> I'll be setting up device maps on the host server to get to the
> filesystems on each VMs virtual disk, and doing read only checks. But
> if anything needs a repair then I could potentially have a fs modified
> on newer kernel & userspace that then has to boot inside a VM with a
> distro with much older kernel and userspace such as Centos 5, which is
> what I'm concerned about.
> 
> Thanks,
> Dan
> --
> To unsubscribe from this list: send the line "unsubscribe linux-ext4" in
> the body of a message to majordomo@vger.kernel.org
> More majordomo info at  http://vger.kernel.org/majordomo-info.html

Re: Using latest e2fsprogs on various offline VM ext filesystems

[Daniel Johnson]

Great, thanks Andreas for confirming that, and for the snapshot tip.
Dan

On 18/04/15 16:12, Andreas Dilger wrote:
> You can safely use newer e2fsprogs on older file systems. This should not
> affect compatibility with older kernels in any way, unless you explicitly
> enable features.
>
> Note that if you are using LVM/DM you could create a snapshot of the
> running device and run e2fsck on that. If no errors are found then you
> don't need to stop the VM at all. You can reset the last checked counter
> and mount count via tune2fs. Only if you find an error in the snapshot do
> you need to unmount the filesystem to fix it.
>
> Cheers, Andreas
>
>> On Apr 18, 2015, at 06:46, Daniel Johnson <dan.u.johns@gmail.com> wrote:
>>
>> Apologies if this "user" question has been asked before.
>>
>> Tomorrow I will be stopping a couple of dozen kvm virtual machines
>> running various Linux distros and running a forced e2fsck on all of
>> their ext filesystems. The distros are a mix of various flavours of
>> Ubuntu, Debian and Centos, so will include some fs created with older
>> kernels and E2fsprogs.
>>
>> The host is 64-bit Centos 6 running vanilla kernel 3.14.39. I figured
>> I'll build the latest E2fsprogs 1.42.12 for this rather than using the
>> Centos 6 e2fsprogs-1.41.12-21.el6.x86_64.
>>
>> My question is; should it be safe in theory to use the latest
>> E2fsprogs on a mix of ext filesystems in this way? I would have
>> thought so but don't want to rely on any assumptions :-)
>>
>> I'll be setting up device maps on the host server to get to the
>> filesystems on each VMs virtual disk, and doing read only checks. But
>> if anything needs a repair then I could potentially have a fs modified
>> on newer kernel & userspace that then has to boot inside a VM with a
>> distro with much older kernel and userspace such as Centos 5, which is
>> what I'm concerned about.
>>
>> Thanks,
>> Dan