* [Drbd-dev] Suggestion: secondary socket address
@ 2007-03-05 19:14 daliv
2007-03-07 0:04 ` Lars Ellenberg
0 siblings, 1 reply; 2+ messages in thread
From: daliv @ 2007-03-05 19:14 UTC (permalink / raw)
To: drbd-dev
suggestion in result of persistent development problems sky2 driver for
"Marvell Technology" network cards:
add possibility to set backup direction of socket for each drbd resource
example :
on host01.domain {
device /dev/drbd0;
disk /dev/sdb1;
address 10.0.2.2:7788;
address-backup 10.0.1.200:7788;
meta-disk internal;
}
or may be with backup leveling
address-backup 10.0.1.200:7788 level 1; // if primary failed -
then here
address-backup 10.0.3.200:7788 level 2; // if primary and
secondary faled - then try to reconnect thru this connection
eth0 : 10.0.1.0/24 - main corporate LAN
eth1 : 10.0.2.0/24 - only interserver network - not for real client
eth2 : 10.0.3.0/24 - additional lan/wan connections
listeners for level 1,2 may be not initialised on startup, but only in
error state of primary listener.
when primary listener will back (periodical recheck) - drbd
communication must return automaticaly to that.
+ additionally notifies about channel changes to admin email.
^ permalink raw reply [flat|nested] 2+ messages in thread
* Re: [Drbd-dev] Suggestion: secondary socket address
2007-03-05 19:14 [Drbd-dev] Suggestion: secondary socket address daliv
@ 2007-03-07 0:04 ` Lars Ellenberg
0 siblings, 0 replies; 2+ messages in thread
From: Lars Ellenberg @ 2007-03-07 0:04 UTC (permalink / raw)
To: drbd-dev
/ 2007-03-05 21:14:48 +0200
\ daliv@apollo.lv:
> suggestion in result of persistent development problems sky2 driver for "Marvell Technology" network cards:
> add possibility to set backup direction of socket for each drbd resource
> example :
>
> on host01.domain {
> device /dev/drbd0;
> disk /dev/sdb1;
> address 10.0.2.2:7788;
> address-backup 10.0.1.200:7788;
> meta-disk internal;
> }
>
> or may be with backup leveling
> address-backup 10.0.1.200:7788 level 1; // if primary failed - then here
> address-backup 10.0.3.200:7788 level 2; // if primary and secondary faled - then try to reconnect thru
> this connection
>
> eth0 : 10.0.1.0/24 - main corporate LAN
> eth1 : 10.0.2.0/24 - only interserver network - not for real client
> eth2 : 10.0.3.0/24 - additional lan/wan connections
I don't see us implementing a fallback
solution like this in module code.
you'd have to monitor and trigger this from userland.
we might add a drbdadm subcommand like
"drbdadm reconnect --use-backup-link=3"
and we might add an optional configuration directive,
so the config would be all in one place.
but you can do something like that right
now already using drbdsetup directly.
btw, if we do something like that, be sure:
we won't do this because some nic driver is not yet stable.
but because it might have general usage to achieve a higher level of
redunancy in the network connection, beyond what is available with the
linux bonding driver.
> listeners for level 1,2 may be not initialised on startup, but only in error state of primary listener.
> when primary listener will back (periodical recheck) - drbd communication must return automaticaly to that.
>
> + additionally notifies about channel changes to admin email.
since this would be triggered from monitoring (likely: scripts),
you may then trigger any other action you like as well.
--
: Lars Ellenberg Tel +43-1-8178292-55 :
: LINBIT Information Technologies GmbH Fax +43-1-8178292-82 :
: Vivenotgasse 48, A-1120 Vienna/Europe http://www.linbit.com :
^ permalink raw reply [flat|nested] 2+ messages in thread
end of thread, other threads:[~2007-03-07 0:04 UTC | newest]
Thread overview: 2+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2007-03-05 19:14 [Drbd-dev] Suggestion: secondary socket address daliv
2007-03-07 0:04 ` Lars Ellenberg
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.