[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: management tools?



On 01/06/2013 01:37 AM, Miles Fidelman wrote:
> A full cloud installation is overkill, and most of the stacks are
> oriented toward having separate storage and compute nodes (our boxes are
> each server + 4 large drives).

With Openstack, you can have each box to serve as multiple roles, like
compute + storage (eg: nova-compute + cinder). I guess that Cinder could
work with DRDB as well, since what it does is only serving some LVM
partitions as iSCSI devices.

> Can you say a little more about DTC as to:
> - general maturity, stability, community support

It's a 10 years old project, which I've been running in production for
as long as the project exists. Therefor, I can't afford to have it just
crash.

There's currently 3 or 4 regular contributors sending patches to the
development list regularly, but they are mostly working on the shared
hosting side. The only problem is that I've been quite busy working on
Openstack over the last months, but I really hope to get back to work on
DTC soon.

> - how helpful is it in keeping track of IP addresses and DNS records
> (I've found keeping track of them to be the most annoying part of
> provisioning VMs)

I think this part is quite ok. There's an IP address manager, on which
you add IP address pools, then for each VM server, you add IPs for new
VMs, and when you add one VM, that IP is marked as used. It isn't
possible to do a mistake, eg an IP cannot be used twice. Finally,
there's a screen which shows the overall IP usage of an IP pool. There
is no new functionality that I wish to add there, I'm satisfied with it.
DTC also does reverse DNS.

> - how helpful is it in terms of configuring, managing, and monitoring
> disks & filesystems (the other pain in the rear when setting up and
> tearing down VMs)

DTC uses Xen, so it will create a new LVM partition for each new VM. It
also has graphs of I/O usage in near real time (eg: max 2 minutes delay
on the RRD graphs), so you can tell if a VM is using too much I/O. It
does that as well for CPU and network usage. There's also an overall
screen which will show a view of all servers at once.

> - does it know anything about pacemaker?

Simple: there's absolutely nothing for HA currently. This would have to
be added. I'd be very happy to receive and review patches though.

I hope that helps,
Cheers,

Thomas

P.S: Please don't CC me, I'm registered to the list.


Reply to: