[DRBD-user] mirror config questions for manual failover

Lars Ellenberg Lars.Ellenberg at linbit.com
Mon Jan 19 13:47:23 CET 2004

Note: "permalinks" may not be as permanent as we would like,
direct links of old sources may well be a few messages off.


/ 2004-01-18 22:18:51 -0500
\ george young:

> There is a private 100Mbit ethernet between the two nodes.
> I use the HA-Linux "IPaddr" script, but heatbeat is *not* enabled.
> 
> Here's my script for pig-app to grab the /db filesystem from pig-db:
> ------------------------------------------------------------
> if ping -c 1 pig-db; then
>     rsh -n pig-db /usr/local/etc/ha.d/resource.d/datadisk drbd_db stop &
>     sleep 30
> fi
> /sbin/drbdsetup /dev/nb1 disconnect
> /sbin/drbdsetup /dev/nb1 net 10.0.0.115:7789 10.0.0.114:7789 C
> /usr/local/etc/ha.d/resource.d/datadisk drbd_db start

please, try to describe in plain english what you think it does.

If I got it right, admin decides: ok, pig-app should now serve the
pig-db filesystem, too.
[ unfortunate choice of names. don't name your nodes like states
[ or resources. that *will* confuse some people, maybe even
[ yourself (after a 36-hour-seventeen-liter-coffee shift)

so you ping the node, if it seems alive, "datadisk stop" remotely,
sleep for a while.
in any case, you "disconnect" (why??)
then newly setup the net      (why??)
and "datadisk start" locally...

there are no "services" you want to stop and start?

> I am also frustrated that it takes 2 hours to sync 36G over a 100Mbit
> private net.  That's a rate of about 5 Mbytes/sec.  Disks on both hosts
> are fast hardware raids.  Am I missing something?

100MBit are at most 12,5 MegaByte per second, minus TCP and DRBD
protocoll overhead, minus conncurrency retries when you have
application access on a syncing device.

so you should see more like about one hour to sync 36G.
ah, and what I see from your drbd.conf,
(though this is guessing) c0d0p5 and c0d0p3 may share the same
physical device, so your sync performance is decreased by seek
time of concurrent access...

add a sync-group = 1 to net { } of drbd_home, and sync-group = 2
to drbd_db, maybe this helps.


> Below is my(common) drbd.conf:
> ------------------------------------------------------------
> resource drbd_home {
>   protocol = C
>   fsckcmd  = fsck -p -y

    as Tony Tay mentioned, for reiser better use /bin/true

>   disk {
>     disk-size = 36707364k
>   }
>   net {
>     sync-min    = 500k

maybe you want to up this.

>     sync-max    = 100M    # maximal average syncer bandwidth

you won't ever see it on 100MBit ...
this basically says: no limitk but don't try too hard.
maybe
	sync-nice = -15
helps, too.


	Lars Ellenberg



More information about the drbd-user mailing list