[DRBD-user] drbd not working with high mtu?

Digimer lists at alteeve.ca
Sat Feb 1 18:20:59 CET 2014

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


On 01/02/14 08:47 AM, Dan Barker wrote:
>> -----Original Message-----
>> From: drbd-user-bounces at lists.linbit.com [mailto:drbd-user-
>> bounces at lists.linbit.com] On Behalf Of Harka Gyozo SA
>> Sent: Saturday, February 01, 2014 1:32 AM
>> To: drbd-user at lists.linbit.com
>> Subject: [DRBD-user] drbd not working with high mtu?
>>
>> Hi!
>>
>> Anyone knows about an issue with mtu?
>> I tried to increase it to max ( there are two nics with a crossover cable,
>> so
>> no switch, router etc.. ). I found that max mtu supported by my card's
>> driver
>> is 7200.
>> ssh, ping (also with big packages), nfs works.
>>
>> drbd keeps saying:] block drbd1: [drbd1_worker/16675] sock_sendmsg time
>> expired, ko = ...
>>
>> If I decrease the mtu to 6000, error message stops, and I can see the
>> syncing
>> in the /proc/drbd (While the mtu is high it's stalled).
>>
>> If I increase the mtu to 6200, kernel message returns, communication
>> stalled.
>>
>> This is a bug? Or maybe I should increase buffer sizes in the config?
>> [ version: 8.3.7 (api:88/proto:86-91) ]
>>
>>
>> HARKA Győző
>> -----------
>
> ssh, ping, nfs all will fragment packets.
>
> Try ping with a 10K size and the don't fragment switch set. Lower the size until it works.
>
> ping -M do -s 6800 -c 1 <host>
>
> Dan

Exactly what I was going to say. Remember though to set the ping payload 
to 28 bytes less than the MTU you set to allow space for the ICMP 
overhead. So if you set 7200 (sounds like a realtek...), use '-s 7172', etc.

-- 
Digimer
Papers and Projects: https://alteeve.ca/w/
What if the cure for cancer is trapped in the mind of a person without 
access to education?



More information about the drbd-user mailing list