Hiya
I hoping someone would be kind enough to help me understand this or
least provide a follow back answer.
Someone asked me yestersday What is the added advantage of using DRBD
with a clustered filesystem (e.g. OCFS), as opposed to using just a
clustered filesystem . I didnt have the
On 05/20/2010 09:53 AM, Brent Clark wrote:
Hiya
I hoping someone would be kind enough to help me understand this or
least provide a follow back answer.
Someone asked me yestersday What is the added advantage of using DRBD
with a clustered filesystem (e.g. OCFS), as opposed to using just a
Hello Brent,
with drbd stored data exists replicated (the 'r' in drbd is for
replicated) in all the nodes of the cluster and that is arguably the
central aim of drbd: to have data consistently replicated on each node.
in typical clustered filesystems scenarios you'dd have many nodes
Comes with Pacemaker:
/usr/lib/ocf/resource.d/heartbeat/LVM
I ended up building RPMs for Pacemaker/Heartbeat. All of the Heartbeat
RPMs I could find in repositories (for example: epel) were too old for
me. If you use x86_64, I can send them to you. Otherwise, the guide is
linked below.
Hi,
Admittedly not a DRBD issue per se, but I guess this list represents quite
some experience in the area: I have two gigabit NICs bonded in balance-rr mode
for DRBD sync. They are directly linked (no switch) to the other other pair in
the other DRBD node.
Before syncing things I was
Do you need multiple destination IPs to properly balance? It is my
understanding that a single stream will only traverse a single link.
Hence the MPIO requirement.
On 5/20/2010 1:07 PM, Bart Coninckx wrote:
Hi,
Admittedly not a DRBD issue per se, but I guess this list represents quite
On Thu, May 20, 2010 at 2:15 PM, Lee Riemer lrie...@bestline.net wrote:
Do you need multiple destination IPs to properly balance? It is my
understanding that a single stream will only traverse a single link. Hence
the MPIO requirement.
Actually, balance-rr is the only mode that can stripe a
On Thursday 20 May 2010 20:39:40 Ben Timby wrote:
On Thu, May 20, 2010 at 2:15 PM, Lee Riemer lrie...@bestline.net wrote:
Do you need multiple destination IPs to properly balance? It is my
understanding that a single stream will only traverse a single link.
Hence the MPIO requirement.
My point was that balance-rr is the ONLY mode to provide multiple link
aggregation for a single tcp stream. That said, the switch's
cooperation is only necessary when a switch is present. The
documentation spells out why.
As both interfaces share the same MAC address, the switch will
recognize
On Thu, May 20, 2010 at 3:43 PM, Bernd Schubert
bs_li...@aakef.fastmail.fm wrote:
On Thursday 20 May 2010, Ben Timby wrote:
Comes with Pacemaker:
/usr/lib/ocf/resource.d/heartbeat/LVM
I ended up building RPMs for Pacemaker/Heartbeat. All of the Heartbeat
RPMs I could find in repositories
On Thursday 20 May 2010 21:26:05 Ben Timby wrote:
My point was that balance-rr is the ONLY mode to provide multiple link
aggregation for a single tcp stream. That said, the switch's
cooperation is only necessary when a switch is present. The
documentation spells out why.
As both interfaces
On Thursday 20 May 2010, Ben Timby wrote:
Comes with Pacemaker:
/usr/lib/ocf/resource.d/heartbeat/LVM
I ended up building RPMs for Pacemaker/Heartbeat. All of the Heartbeat
RPMs I could find in repositories (for example: epel) were too old for
me. If you use x86_64, I can send them to
On Thu, May 20, 2010 at 3:52 PM, Bart Coninckx bart.conin...@telenet.be wrote:
Since testing with a switch resulted in even poorer results, I'll shy away
from that for now. Will have a go at investigating with Wireshark though
seeing the problem will probably not indicate a solution right of
On Thu, May 20, 2010 9:43 pm, Michael wrote:
Send me please this part:
cibadmin -Q | grep node_state
cibadmin -Q | grep node_state node_state id=FXVS2 uname=FXVS2
Oh, uppercased hostnames. This is a known bug and it has already been
fixed. I've uploaded a new dmctest-0.7.4.dev.1.jar to
- Original message -
Since testing with a switch resulted in even poorer results, I'll shy away
from that for now.
You might use a better switch, the 1810's are web based mgmt only and have
small buffers iirc and are not very enterprisable units, their performance
might not be
15 matches
Mail list logo
