Hi Lucian!
I can and will only speak about question#1. If I correctly read your
past mails, you're trying to implement DRBD for some VMware platform,
right ? and then your actual question is "How can I DRBDize only
selected VM's vmdk files, not all VMs ?" right ?. If so, IMHO I believe
the immediate answer is "No", AFAIK there's no mean to implement DRBD at
the vSphere hypervisor level, even a Service Console, which is what you
would need to make it work. However I see 2 ways that would provide you
with kind of a solution :
1) You can implement DRBD within your VMs. This does not sound the best
solution as it will require an equivalent "secondary VM" on the other
side, plus you will have to manage replication traffic within both VMs :
Complicated, and personnally I figure out data replication as an
infrastructure task, not an application task.
2) You can give Raw Device Mappings to your VMs instead of regular VMDK
files. In such case, you woud present one or several LUNs from your DRBD
storage platform to your VM via iSCSI for instance. RDM will make them
appear just like if they were regular VMDK files, except it's not a
file, it's a LUN formatted with whatever format you require. Then, it's
no big deal to replicate such a LUN using standard and supported DRBD ways.
However, if you chose solution 2), you could ask yourself if, while
replicating some LUNs, it wouldn't worth replicate a complete VMFS LUN
storing VMDKs, rather classic LUNs containing ext3 or something. Unless
replicated volume size is too big, because of your replication network
bandwitdh or something, when your replicating, why not replicating
everything ? Well, this choice is up to you...
Hope this helps.
Best regards,
Pascal.
Le 12/07/2011 10:51, Lucian Torje a écrit :
Hello,
I'm trying to configure DRBD & Pacemaker & Corosync & RabbitMQ and I
need some help in the following areas:
1) Is it possible to configure to use a file containing a filesystem
in DRBD ( filesystem within another filesystem file) ?
2) Pacemaker automatically mounts the drdb device on the master node
(the cluster contains a primary-secondary nodes) while the slave
doesn't have access in order not to corrupt the data, right?
I have the following configuration and the filesystem was not mounted
for some reason (although DRBD management console displays that
drbd:r0 block device is synchronized and configured as primary-secodary):
$crm configure show
node ha-node1
node ha-node2
primitive bunny ocf:rabbitmq:rabbitmq-server \
params mnesia_base="/mnt/drbd1"
primitive drbd ocf:linbit:drbd \
params drbd_resource="drbd1" \
op monitor interval="60s" \
meta is-managed="true"
primitive drbd_fs ocf:heartbeat:Filesystem \
params device="/dev/drbd0" directory="/mnt/drbd1" fstype="ext3"
ms drbd_ms drbd \
meta master-max="1" master-node-max="1" clone-max="2"
clone-node-max="1" notify="true"
colocation bunny_on_fs inf: bunny drbd_fs
colocation fs_on_drbd inf: drbd_fs drbd_ms:Master
order bunny_after_fs inf: drbd_fs bunny
order fs_after_drbd inf: drbd_ms:promote drbd_fs:start
property $id="cib-bootstrap-options" \
dc-version="1.0.8-042548a451fce8400660f6031f4da6f0223dd5dd" \
cluster-infrastructure="openais" \
expected-quorum-votes="2" \
last-lrm-refresh="1310458848"
$ crm_mon
============
Last updated: Tue Jul 12 01:30:55 2011
Stack: openais
Current DC: ha-node1 - partition with quorum
Version: 1.0.8-042548a451fce8400660f6031f4da6f0223dd5dd
2 Nodes configured, 2 expected votes
3 Resources configured.
============
Online: [ ha-node1 ha-node2 ]
3) How can I mount a shared folder instead of DRDB just to test the
rest of configuration?
Thank you
Best regards,
Lucian-Paul Torje
_______________________________________________
drbd-user mailing list
[email protected]
http://lists.linbit.com/mailman/listinfo/drbd-user
_______________________________________________
drbd-user mailing list
[email protected]
http://lists.linbit.com/mailman/listinfo/drbd-user
