[DRBD-user] DRBD (pingACK did not arrived in time issue)

Felix Frank ff at mpexnet.de
Thu Apr 21 12:21:11 CEST 2011

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


On 04/19/2011 05:38 PM, Aditya bajaj wrote:
> Hello all,
> 
> we have a two node cluster setup  , under heavy load (created using many
> paralallel ping flood sessions)  the following messages are being
> observed in syslog.
> 
> Apr 19 09:49:59 err CLA-0 kernel: drbd6: PingAck did not arrive in time

<snip>

> _this_host {
>         device                  "/dev/drbd0";
>         disk                    "/dev/VG_62/sysimg";
>         meta-disk               internal;
>         address                 192.168.0.4:49156
> <http://192.168.0.4:49156>;
> }
> _remote_host {
>         address                 192.168.0.5:49156
> <http://192.168.0.5:49156>;
> }

Hi,

am I right to assume that DRBD uses each server's single NIC?

You could probably do better by adding a NIC to each server, connect
them back-to-back (crossover cable) and use that link for DRBD only.

As an aside, was it customary with 8.0 to have assymetric configurations
on the nodes? You may want to upgrade to 8.3 either way ;-)

HTH,
Felix



More information about the drbd-user mailing list