[DRBD-user] 10Gb ethernet ?

Igor Neves igor at 3gnt.net
Fri Jun 20 12:17:22 CEST 2008

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


Lee Christie wrote:
> Again, I believe that in 802.3ad when transmitting to the same 
> destination, packets will go over the same link.

We use always 802.3ad, but the problem it's not the 802.3ad, the problem 
it's the algorithm switch uses, for distribution traffic.

Linux in mode 0, balance-rr can distribute traffic for the same 
destination mac over multiple links. But the problem it's the switch on 
is side does not distribute traffic the same way linux does. Maybe with 
the trick of multiple vlan's the switch will do it.

We normally don't work with cisco, we tried this with linksys and we 
don't manage to get full speed of 2 our 4 NIC Gbit Ethernet.

We never get the trunk to work at full speed of the multiple NIC's, with 
one switch in the midle, for the same destination. If someone have this 
working, please tell me with switch and witch modes you use.

>  
> I think we tested ftp throughput between two servers (connected via a 
> cisco switch) using the various bonding methods to see if any 
> performance gains were there. I think the answer was no because it was 
> point to point however I note with interest the snippet below which 
> suggests the two different ports need to be in different vlans. Not 
> quite sure why that would be but I'll take it on face value :)

I suggest you always test network stuff with iperf. Ftp you are taking 
in the game too many stuff, like disk and kernel caches.
Iperf uses directly the NIC, from memory to memory, the limit will be 
your your network hardware you are trying to test.

>  
> The bottom line is that no aggregation of multiple "slow" links will 
> ever beat a single "fast" link. And there are 
> compromises/caveats/complexity along the way.
>  
> We use link bonding for resilience across cisco 3750 stacks as two 
> switches can be treated as a single logical entity and when you drop a 
> switch or link there is almost no packet loss.
>  
> 10Gb-e is getting cheaper all the time. Intel are currently offering a 
> 2 for 1 deal on the 10Gb-e adapters we bought in for testing, so we 
> paid 700 GBP for a pair. Configuration was a simple matter of dropping 
> in the cards, building the kernel module and a wee bit of 
> modprobe/kudzu , configure of eth2 and bingo - in business.
>  
>  
>
>     ------------------------------------------------------------------------
>     *From:* drbd-user-bounces at lists.linbit.com
>     [mailto:drbd-user-bounces at lists.linbit.com] *On Behalf Of *Sören
>     Malchow
>     *Sent:* 19 June 2008 17:05
>     *To:* Ralf Gross
>     *Cc:* drbd-user-bounces at lists.linbit.com; drbd-user at lists.linbit.com
>     *Subject:* Re: [DRBD-user] 10Gb ethernet ?
>
>
>     Hi,
>
>     i am sorry if i miss the point, but i did not read the older
>     messages.
>
>     But why not use 802.3ad aka dynamic link aggregation, we have that
>     successfully configured with HP and Nortel switches ( also with
>     machines running drbd over those links ).
>
>     As far as i know cisco switches also support 802.3ad.
>
>     Regards
>     Soeren
>
>
>
>
>     *Ralf Gross <Ralf-Lists at ralfgross.de>*
>     Sent by: drbd-user-bounces at lists.linbit.com
>
>     19.06.2008 17:49
>
>     	
>     To
>     	drbd-user at lists.linbit.com
>     cc
>     	
>     Subject
>     	Re: [DRBD-user] 10Gb ethernet ?
>
>
>
>     	
>
>
>
>
>
>     Lars Ellenberg schrieb:
>     > On Thu, Jun 19, 2008 at 01:03:11PM +0100, Lee Christie wrote:
>     > > In any event, I'm no expert on channel bonding, but in a 2-server
>     > > configuration, where the Ips and MAC addresses are fixed at
>     either end,
>     > > how can you use all 4 channels ? I was always under the
>     impression that
>     > > the bonding used an algorithm based on src/dest IP/Mac to
>     choose which
>     > > link to send data down, so in a point to point config it would
>     always be
>     > > the same link.
>     >
>     > "balance-rr" aka mode 0 for linux bonding schedules packets
>     round robin
>     > over the available links.
>
>     balance-rr will not help in a cisco environment, because the switch
>     will still use the same ports and does not perform rr load balancing.
>     The only way to get rr working with cisco switches was to use 2
>     different vlans.
>
>     eth0 <---- vlan x ----> eth0
>     eth1 <---- vlan y ----> eth1
>
>     At least this is my experience and our CCNP's (or what they are
>     called) told me the same.
>
>     With the above trick I was able to get ~1.6x GbE throughput with
>     the netpipe benchmark (after tuning the reorder kernel parameter). I
>     didn't used the conenction for drbd, I tried to speed up our backup.
>     But the funny thing was, it slowed down. Even the ftpd or samba was
>     slower over this link as it was with the xor mode or just one GbE NIC.
>
>     > but still, for a single tcp connection, given some tcp_reorder
>     tuning,
>     > the strong gain you get from 2x 1GbE (1.6 to 1.8 * that of one
>     channel)
>     > degrades again to effectively less than one channel if you try
>     to use 4x.
>     >
>     > again, "more" is not always "better".
>     > for the usage pattern of drbd (single tcp connection with bulk
>     data) the
>     > throughput-optimum linux bonding seems to be 2x, with 3x you are
>     back to
>     > around the same throughput as 1x, with 4x you are even worse
>     than 1x,
>     > because packet reordering over bonded GbE and tcp congestion control
>     > don't work well together for sinlge tcp links.
>
>     Very true.
>
>     Ralf
>     _______________________________________________
>     drbd-user mailing list
>     drbd-user at lists.linbit.com
>     http://lists.linbit.com/mailman/listinfo/drbd-user
>
> ------------------------------------------------------------------------
>
> This email may contain legally privileged and/or confidential 
> information. It is solely for and is confidential for use by the 
> addressee. Unauthorised recipients must preserve, observe and respect 
> this confidentiality. If you have received it in error please notify 
> us and delete it from your computer. Do not discuss, distribute or 
> otherwise copy it.
>
> Unless expressly stated to the contrary this e-mail is not intended 
> to, and shall not, have any contractually binding effect on the 
> Company and its clients. We accept no liability for any reliance 
> placed on this e-mail other than to the intended recipient. If the 
> content is not about the business of this Company or its clients then 
> the message is neither from nor sanctioned by the Company.
>
> We accept no liability or responsibility for any changes made to this 
> e-mail after it was sent or any viruses transmitted through this 
> e-mail or any attachment. It is your responsibility to satisfy 
> yourself that this e-mail or any attachment is free from viruses and 
> can be opened without harm to your systems.
>
>  
>
> ------------------------------------------------------------------------
>
> _______________________________________________
> drbd-user mailing list
> drbd-user at lists.linbit.com
> http://lists.linbit.com/mailman/listinfo/drbd-user
>   


-- 
Igor Neves <igor.neves at 3gnt.net>
3GNTW - Tecnologias de Informação, Lda
 
 SIP/JID/MSN: igor at 3gnt.net 
 Mobile Phone: 00351914503611
 Local Phone: 00351252377120





More information about the drbd-user mailing list