[DRBD-user] Slower disk throughput on DRBD partition

Felix Frank ff at mpexnet.de
Wed Feb 1 09:27:47 CET 2012

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


Hi,

On 01/31/2012 06:03 PM, Frederic DeMarcy wrote:
> No replication over the 1Gb/s crossover cable is taking place since the secondary node is down yet there's x2 lower
> disk performance.

the *throughput* is down. For MySQL, you should by all means be more
concerned about the latency.

Be that as it may, halved throughput seems rather severe. Please bear in
mind though: Writing to a disconnected DRBD is not implicitly cheaper
than in the connected case. For one thing, DRBD needs to update the
quicksync bitmap as you're writing unreplicated data. This may cost you
throughput-wise (still, this shouldn't explain a factor 2).

What's your network capable of? What's your maximum throughput in the
connected case?

Cheers,
Felix



More information about the drbd-user mailing list