[DRBD-user] Write performance again

Roland Kaeser roland.kaeser at ziil.ch
Wed Sep 26 16:18:34 CEST 2012

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

Hello Sebastian 

Thanks for the hint. The be sure that the settings are exactly the same on both sides, I wrote a script to set these settings by ssh on both nodes equally. So the current value for max_sectors_kb is: 64 on both nodes. But I played around with these values and nothing helped. When the nodes are connected, the write speed is even lower than in single node mode. I think there must be some kind of bottleneck between the blocksizes / ioscheduler, etc. and drbd but could find the igniting hint. 

Strange is also that the sync speed is very low (triggered manually resync to retest that) 66MB /s. All tests with different config settings showed no performance changement. blktrace also shows nothing more flashy which could lead to more. while resyncing and dd'ing the iostat shows 1.07% iowat which is imho very low value for a system under full io load. 

Is there a possibility to debug the network usage and internal network stats of drbd besides drbdsetup show 0? 



----- Ursprüngliche Mail -----
Von: "Sebastian Riemer" <sebastian.riemer at profitbricks.com> 
An: "Roland Kaeser" <roland.kaeser at ziil.ch> 
CC: drbd-user at lists.linbit.com 
Gesendet: Mittwoch, 26. September 2012 15:49:07 
Betreff: Re: [DRBD-user] Write performance again 

Hi Roland, 

I've analysed the DRBD 8.4.1 kernel code some time ago. 

DRBD has a dynamic IO request size detection which always starts at 4 
KiB. If you've got a connection to the other system, then it negotiates 
a higher IO request size (128 KiB because of hashing functions). If you 
loose the connection, it's 4 KiB again. 

You can use the tool "blktrace" to trace how big the IO requests are. Or 
you can look at /sys/block/drbd0/queue/max_sectors_kb. 

IMHO this must become settable to the limits from the devices below if 
you already know that "max_sectors_kb" is completely equal on both sides. 



On 26.09.2012 14:50, Roland Kaeser wrote: 
> Only the drbd is very slow. (Single node, just to test write 
> perfomance on local block device. 


Ziil Informatiklösungen GmbH 
Roland Käser 
Leiter Technik 
Romanshornerstr. 134, 
CH-8280 Kreuzlingen 
Tel: +41 71 671 27 78 
Fax: +41 71 671 27 79 
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.linbit.com/pipermail/drbd-user/attachments/20120926/17dc0e1c/attachment.htm>

More information about the drbd-user mailing list