DART 5.6.49.3 with 2xdatamover- how configure nics and devices for fast response in CIFS

I would use option 2 (both fsn and trunking) if I had two switches and cared that my CIFS server is still working if one switch fails

If you don't have two switches or all your clients are connected through one switch it's a moot point

you won't find much on Powerlink how to configure your switch - you really need to look at your switch vendors docs for EtherChannel or LACP - it's pretty straightforward

Rainer

i did FSN that consists of LACP (primary connection): cge0,cge1,cge2 and regular cge3.  My position is that i would like to give my users as much connectivity as possible with the caution that if the network switch that has LACP connection dies, they will be running in "degraded" mode on cge3 only.

good choice

in your choice,

how the nics work just 1 primary and the others 3 in standby mode, is not clear to me

other thing making redundancy from the switch side,

how could you connect them with 2 ethernet switches,

is it necesary work with 2 switches yes or no and why?

any sugestion

could be this work?

standby datamover  ---------------------      primary datamover

cge0 ----- cge1 ----- cge2 ----- cge3       cge0 ------ cge1 ------ cge2 ------ cge3

  |              |             |               |            |               |              |             |

sw2p1    sw1p1     sw2p3     sw1p3     sw1p1      sw2p2     sw1p3       s2wp3

how could you configure FSN and LACP with two switches is it posible yes or not.

thank you

Maynor

questions:

when configuring LACP on every port that is going to be used

the first three ports cge0,cge1,cge2 have traffic at the same time?

converting this connection in a 3Gb interface is it correct?

other thing following your configuration

you mean that at the switch2 side, there is going to be only 2 conections coming from primary and standby datamovers?

thank you

mc

questions: when configuring LACP on every port that is going to be used the first three ports cge0,cge1,cge2 have traffic at the same time? converting this connection in a 3Gb interface is it correct?

yes, they will have traffic at the same time but think of it as a 3 x 1G lane autobahn. No one car can drive at 3G but 3 cars can drive at 1G concurrently .

other thing following your configuration you mean that at the switch2 side, there is going to be only 2 conections coming from primary and standby datamovers? thank you mc

yep, 6 connections on switch 1 and 2 connections on switch 2.

thnks you

now see about this new config i am thinking

from primary datamover all 4 port configured in LACP mode connected to switch-1 and

from standby datamover all 4 port connected to switch-2, all of this, in order of getting the most performance connectivity,

is it valid?

what could be the risks?

my questions are in this new config:

1. if one port from a  primary datamover or one port used for LACP of switch-1 gets failed, does the standby DM takes personality of primary  DM?

2. what could happend if entire switch-1 failed, does the standby datamover takes personality of primary DM too.

thanks

network failure is NOT a trigger for automatic failover

yesterday, following steps to configure LACP on my existing NAS, in order to get more performance.

from server_2 datamover my ports were connected to a cisco 3560g switch.

firts before changes, i have a CIFS server with a VDM, with a fsn device like this (cge0-primary, cge1-standby)

i test my conection tranfering  a cd image of 725 Megas of size, and the transfer were made in 1:15 minutos

from a client computer to a share folder

then i reconfigure the conection, to LACP with and interface like this (lacp01=cge0,cge1,cge2), and configuring the switch with

a etherchannel lacp group1 (including 3 ports, to connect cge0,cge1,cgd2).

With all reconfigure done.

i made the same tranfers test, but it takes de same time, !!!!! whats the reason, or is there something wrong.

i spect at least minize the tranfer to 30 seconds. but unfortunately it was not so 

any idea.

mc

you have one tcp connection, meaning you are only using one of those 1G lanes.  Try the same test with 3 different workstations.