[DRBD-user] ProtocollError using on-congestion pull-ahead

Lars Ellenberg lars.ellenberg at linbit.com
Fri Feb 8 11:04:39 CET 2013

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


On Fri, Feb 08, 2013 at 10:41:19AM +0100, Holger Steinhaus wrote:
> >>         resync-rate 200M;
> >Uhm, what? That's more than your link can do. You want to max this at
> >30% of your maximum throughput, tops. Otherwise the resync may interfere
> >with on-line write performance.
> >
> >>         al-extents 389;
> >Why? This should be chosen to be as large (a prime number) as possible.
> 
> Sorry, remainders of some experiments to overcome the trouble (idea
> was to de-limit synchronization bandwith and to limit the number of
> hot extents). I just return them to their defaults (removed
> assignmend) and tried again - same result. In a second experment, I
> limited the resync to 20M and increased the al extends to 3389 as
> stated in the doc. There is absolutely no change in behavior - still
> lots of ProtocolErrors.
> 
> I have almost forgotten the tuning approach for now - I am much more
> scared about stability and data integrity due to the obvious
> protocol desynchronizations...

You ever heard of *logs* ?
 ;-)

Maybe post some ...


-- 
: Lars Ellenberg
: LINBIT | Your Way to High Availability
: DRBD/HA support and consulting http://www.linbit.com

DRBD® and LINBIT® are registered trademarks of LINBIT, Austria.
__
please don't Cc me, but send to list   --   I'm subscribed



More information about the drbd-user mailing list