Re: [gentoo-user] Best filesystem for virtualized gentoo mail server - WAS: vmWare HowTo / best practices

[J. Roeleveld]

On Sat, April 20, 2013 17:33, Alan McKinnon wrote:
> On 20/04/2013 17:00, Tanstaafl wrote:
>> Thanks for the responses so far...
>>
>> Another question - are there any caveats as to which filesystem to use
>> for a mail server, for virtualized systems? Ir do the same
>> issues/questions apply (ie, does the fact that it is virtualized not
>> change anything)?
>>
>> If there are none, I'm curious what others prefer.
>>
>> I've been using reiserfs on my old mail server since it was first set up
>> (over 8 years ago). I have had no issues with it whatsoever, and even
>> had one scare with a bad UPS causing the system to experienc an unclean
>> shutdown - but it came back up, auto fsck'd, and there was no 'apparent'
>> data loss (this was a very long time ago, so if there had been any
>> serious problems, I'd have known about it long go).
>>
>> I've been considering using XFS, but have never used it before.
>>
>> So, anyway, opinions are welcome...
>
>
> Virtualization can change things, and it's not really intuitive.
>
> Regardless of what optimizations you apply to the VM, and regardless of
> what kind of virtualization is in use on the host, you are still going
> to be bound by the disk and fs behaviour of the host. If VMWare gives
> you a really shitty host driver, then something really shitty is going
> to be the best you can achieve.

This can be improved by not using file-backed disk devices.
Not sure if ESXi can do this. XenServer supports LVM-backed disk devices,
which reduces the overhead for the FS significantly.

> Disks aren't like eg NICs, you can't easily virtualize them and give the
> guest exclusive access in the style of para-virtualization (I can't
> imagine how that would even be done).

There is one option for this, but then you need to push the whole
disk-device for the VM. In other words you pass "sda" to the VM, instead
of a file or partition.
To avoid lousy I/O driver on the host, you could try passing the
disk-controller directly to the VM. But then the VM has it's own set of
disks that can not be used for other VMs.


> FWIW, I have two mail relays (no mail storage) running old postfix
> versions on FreeBSD. I expected throughput to differ when virtualized on
> ESXi, but in practice I couldn't see a difference at all - maybe the
> mail servers were very under-utilized. Considering this pair deal with
> anything between 500,000 to a million mails a day total, I would not
> have considered them "under-utilized". Just goes to show how opinions
> are often worthless but numbers buys the whiskey :-)

If Postfix only passes emails through, then it only uses the mail-spool
for temporary storage. For that, it doesn't require a lot of disk I/O.
Most filtering is done in memory, afaik.

My experience with ESXi is that it can have issues with networking when
the versions of the host don't match perfectly or the time is not synched
correctly or VMs are moved around a lot by an Admin that likes to play
with that...
Doing large multi-system installations on an ESXi cluster while the VMs
are moved around can occasionally fail because of a bad network-layer.

--
Joost

>
>
> --
> Alan McKinnon
> alan.mckinnon@×××××.com
>
>
>
>