drbd (was nmbd)

Philipp Reisner kde at ist.org
Mon Nov 8 23:54:25 MST 1999


Am Mon, 08 Nov 1999 schriebst Du:
>In article <14373.47175.80544.454768 at dukat.scot.redhat.com>, Stephen C. Tweedie <sct at redhat.com> wrote:
>
>> On Sat, 6 Nov 1999 12:13:42 +0100, Philipp Reisner <kde at ist.org> said:
>>> If the failed node comes up again, it is a new secondary node and has
>>> to synchronise its content to the primary. This, of course, will happen
>>> whithout interruption of service in the background. 
>
>> This is the cluster partition problem, and it is one of the hardest
>> parts of HA.
>
>IBM solves it by letting the secondary site call the primary one via a
>dedicated independent phone line. Seems pretty straightforward, and needs
>to be provided by the HA software, not the network block device.
>
>I'd be interested how drbd handles the proper packet sequence, though. 
>

For now drbd uses only one TCP connection, thus the packets arrive
in order.

--Philipp
------------------------------------------------------------------------
     Want to try something new?            Are you a Linux hacker?
		    Volunteer in testing mergemem!
	      (Get it from http://das.ist.org/mergemem)
-----
  Philipp Reisner                PGP: http://der.ist.org/~kde/pgp.asc



More information about the Linux-HA mailing list