Mailing List Archive: LACP with Foundry and Linux Machine

LACP with Foundry and Linux Machine

Mar 27, 2014, 10:44 AM

Post #1 of 18 (8628 views)

Hi All,

I have been trying LACP between Foundry and Linux machine since some
days but not able to figure out the issue and resolve the issue. I
didn't see much configurations thats needs to be done on foundry. I have
configured my linux box on mode 0 and bonding seems to be up but still I
am not able to pass traffic through it.

Serveriron Configurations:

interface ethernet 4/2
port-name **LACP-with-Linux***
no spanning-tree
link-aggregate active
!
interface ethernet 4/3
port-name **LACP-with-Linux***
no spanning-tree
link-aggregate active
!

SI#show version
SW: Version 09.3.00eTD4 Copyright (c) 1996-2003 Foundry Networks, Inc.
Compiled on Jun 28 2005 at 18:57:32 labeled as WXR09300e
HW: ServerIronGT E-2 Router, SYSIF version 21, Serial #: Non-exist
==========================================================================
SL 1: B0GMR WSM2 Management Module, SYSIF 2, M6, ACTIVE
Serial #: CH39040033
0 MB SHM, 2 Application Processors
16384 KB BRAM, SMC version 1, BM version 21
SW: (1)09.3.00eTF2 (2)09.3.00eTF2
==========================================================================
SL 4: J-BxGC16 JetCore Gig Copper Module, SYSIF 2
Serial #: CH33060779
4096 KB BRAM, JetCore ASIC IGC version 49, BIA version 8a
32768 KB PRAM and 2M-Bit*1 CAM for IGC 12, version 0449
32768 KB PRAM and 2M-Bit*1 CAM for IGC 13, version 0449
32768 KB PRAM and 2M-Bit*1 CAM for IGC 14, version 0449
32768 KB PRAM and 2M-Bit*1 CAM for IGC 15, version 0449

SI#show link-aggregate
System ID: 000c.db5d.b066
Long timeout: 90, default: 90
Short timeout: 3, default: 3
Port [Sys P] [Port P] [ Key ] [Act][Tio][Agg][Syn][Col][Dis][Def][Exp][Ope]
4/2 1 1 483 Yes S Agg Syn Col Dis Def No Ope
4/3 1 1 483 Yes S Agg Syn Col Dis Def No Ope

Status of linux machine:
MII Status: up
MII Polling Interval (ms): 30000
Up Delay (ms): 0
Down Delay (ms): 0

Slave Interface: eth0
MII Status: up
Speed: 1000 Mbps
Duplex: full
Link Failure Count: 0
Permanent HW addr: 00:e0:ed:2b:a1:99
Slave queue ID: 0

Slave Interface: eth1
MII Status: up
Speed: 1000 Mbps
Duplex: full
Link Failure Count: 0
Permanent HW addr: 00:e0:ed:2b:a1:98
Slave queue ID: 0

--
Regards,
--------
Rupesh Basnet

_______________________________________________
foundry-nsp mailing list
foundry-nsp@puck.nether.net
http://puck.nether.net/mailman/listinfo/foundry-nsp

Re: LACP with Foundry and Linux Machine [ In reply to ]

nick at foobar

Mar 27, 2014, 11:18 AM

Post #2 of 18 (8541 views)

Permalink

On 27/03/2014 17:44, Rupesh Basnet wrote:
> I have been trying LACP between Foundry and Linux machine since some days
> but not able to figure out the issue and resolve the issue. I didn't see
> much configurations thats needs to be done on foundry. I have configured my
> linux box on mode 0 and bonding seems to be up but still I am not able to
> pass traffic through it.

have you assigned the correct untagged vlan to e4/2 and 4/3?

Nick

_______________________________________________
foundry-nsp mailing list
foundry-nsp@puck.nether.net
http://puck.nether.net/mailman/listinfo/foundry-nsp

Re: LACP with Foundry and Linux Machine [ In reply to ]

georgeb at gmail

Mar 27, 2014, 11:24 AM

Post #3 of 18 (8537 views)

Permalink

"I have configured my linux box on mode 0"

Wouldn't you need to use mode 4?

802.3ad or 4

IEEE 802.3ad Dynamic link aggregation. Creates
aggregation groups that share the same speed and
duplex settings. Utilizes all slaves in the active
aggregator according to the 802.3ad specification.

On Thu, Mar 27, 2014 at 10:44 AM, Rupesh Basnet <brupesh@subisu.net.np>wrote:

> Hi All,
>
> I have been trying LACP between Foundry and Linux machine since some days
> but not able to figure out the issue and resolve the issue. I didn't see
> much configurations thats needs to be done on foundry. I have configured my
> linux box on mode 0 and bonding seems to be up but still I am not able to
> pass traffic through it.
>
> Serveriron Configurations:
>
> interface ethernet 4/2
> port-name **LACP-with-Linux***
> no spanning-tree
> link-aggregate active
> !
> interface ethernet 4/3
> port-name **LACP-with-Linux***
> no spanning-tree
> link-aggregate active
> !
>
>
> SI#show version
> SW: Version 09.3.00eTD4 Copyright (c) 1996-2003 Foundry Networks, Inc.
> Compiled on Jun 28 2005 at 18:57:32 labeled as WXR09300e
> HW: ServerIronGT E-2 Router, SYSIF version 21, Serial #: Non-exist
> ==========================================================================
> SL 1: B0GMR WSM2 Management Module, SYSIF 2, M6, ACTIVE
> Serial #: CH39040033
> 0 MB SHM, 2 Application Processors
> 16384 KB BRAM, SMC version 1, BM version 21
> SW: (1)09.3.00eTF2 (2)09.3.00eTF2
> ==========================================================================
> SL 4: J-BxGC16 JetCore Gig Copper Module, SYSIF 2
> Serial #: CH33060779
> 4096 KB BRAM, JetCore ASIC IGC version 49, BIA version 8a
> 32768 KB PRAM and 2M-Bit*1 CAM for IGC 12, version 0449
> 32768 KB PRAM and 2M-Bit*1 CAM for IGC 13, version 0449
> 32768 KB PRAM and 2M-Bit*1 CAM for IGC 14, version 0449
> 32768 KB PRAM and 2M-Bit*1 CAM for IGC 15, version 0449
>
>
> SI#show link-aggregate
> System ID: 000c.db5d.b066
> Long timeout: 90, default: 90
> Short timeout: 3, default: 3
> Port [Sys P] [Port P] [ Key ] [Act][Tio][Agg][Syn][Col][Dis]
> [Def][Exp][Ope]
> 4/2 1 1 483 Yes S Agg Syn Col Dis Def No Ope
> 4/3 1 1 483 Yes S Agg Syn Col Dis Def No Ope
>
>
> Status of linux machine:
> MII Status: up
> MII Polling Interval (ms): 30000
> Up Delay (ms): 0
> Down Delay (ms): 0
>
> Slave Interface: eth0
> MII Status: up
> Speed: 1000 Mbps
> Duplex: full
> Link Failure Count: 0
> Permanent HW addr: 00:e0:ed:2b:a1:99
> Slave queue ID: 0
>
> Slave Interface: eth1
> MII Status: up
> Speed: 1000 Mbps
> Duplex: full
> Link Failure Count: 0
> Permanent HW addr: 00:e0:ed:2b:a1:98
> Slave queue ID: 0
>
> --
> Regards,
> --------
> Rupesh Basnet
>
> _______________________________________________
> foundry-nsp mailing list
> foundry-nsp@puck.nether.net
> http://puck.nether.net/mailman/listinfo/foundry-nsp
>

Re: LACP with Foundry and Linux Machine [ In reply to ]

brupesh at subisu

Mar 27, 2014, 11:27 AM

Post #4 of 18 (8528 views)

Permalink

Hi George,

First I tried mode 4(802.3ad) at first but LACP didn't formed at all.

Regards,
--------
Rupesh Basnet

On 03/28/2014 12:09 AM, George B. wrote:
> "I have configured my linux box on mode 0"
>
> Wouldn't you need to use mode 4?
>
> 802.3ad or 4
>
> IEEE 802.3ad Dynamic link aggregation. Creates
> aggregation groups that share the same speed and
> duplex settings. Utilizes all slaves in the active
> aggregator according to the 802.3ad specification.
>
>
> On Thu, Mar 27, 2014 at 10:44 AM, Rupesh Basnet <brupesh@subisu.net.np
> <mailto:brupesh@subisu.net.np>> wrote:
>
> Hi All,
>
> I have been trying LACP between Foundry and Linux machine since
> some days but not able to figure out the issue and resolve the
> issue. I didn't see much configurations thats needs to be done on
> foundry. I have configured my linux box on mode 0 and bonding
> seems to be up but still I am not able to pass traffic through it.
>
> Serveriron Configurations:
>
> interface ethernet 4/2
> port-name **LACP-with-Linux***
> no spanning-tree
> link-aggregate active
> !
> interface ethernet 4/3
> port-name **LACP-with-Linux***
> no spanning-tree
> link-aggregate active
> !
>
>
> SI#show version
> SW: Version 09.3.00eTD4 Copyright (c) 1996-2003 Foundry
> Networks, Inc.
> Compiled on Jun 28 2005 at 18:57:32 labeled as WXR09300e
> HW: ServerIronGT E-2 Router, SYSIF version 21, Serial #: Non-exist
> ==========================================================================
> SL 1: B0GMR WSM2 Management Module, SYSIF 2, M6, ACTIVE
> Serial #: CH39040033
> 0 MB SHM, 2 Application Processors
> 16384 KB BRAM, SMC version 1, BM version 21
> SW: (1)09.3.00eTF2 (2)09.3.00eTF2
> ==========================================================================
> SL 4: J-BxGC16 JetCore Gig Copper Module, SYSIF 2
> Serial #: CH33060779
> 4096 KB BRAM, JetCore ASIC IGC version 49, BIA version 8a
> 32768 KB PRAM and 2M-Bit*1 CAM for IGC 12, version 0449
> 32768 KB PRAM and 2M-Bit*1 CAM for IGC 13, version 0449
> 32768 KB PRAM and 2M-Bit*1 CAM for IGC 14, version 0449
> 32768 KB PRAM and 2M-Bit*1 CAM for IGC 15, version 0449
>
>
> SI#show link-aggregate
> System ID: 000c.db5d.b066
> Long timeout: 90, default: 90
> Short timeout: 3, default: 3
> Port [Sys P] [Port P] [ Key ]
> [Act][Tio][Agg][Syn][Col][Dis][Def][Exp][Ope]
> 4/2 1 1 483 Yes S Agg Syn Col Dis Def
> No Ope
> 4/3 1 1 483 Yes S Agg Syn Col Dis Def
> No Ope
>
>
> Status of linux machine:
> MII Status: up
> MII Polling Interval (ms): 30000
> Up Delay (ms): 0
> Down Delay (ms): 0
>
> Slave Interface: eth0
> MII Status: up
> Speed: 1000 Mbps
> Duplex: full
> Link Failure Count: 0
> Permanent HW addr: 00:e0:ed:2b:a1:99
> Slave queue ID: 0
>
> Slave Interface: eth1
> MII Status: up
> Speed: 1000 Mbps
> Duplex: full
> Link Failure Count: 0
> Permanent HW addr: 00:e0:ed:2b:a1:98
> Slave queue ID: 0
>
> --
> Regards,
> --------
> Rupesh Basnet
>
> _______________________________________________
> foundry-nsp mailing list
> foundry-nsp@puck.nether.net <mailto:foundry-nsp@puck.nether.net>
> http://puck.nether.net/mailman/listinfo/foundry-nsp
>
>

Re: LACP with Foundry and Linux Machine [ In reply to ]

georgeb at gmail

Mar 27, 2014, 11:32 AM

Post #5 of 18 (8534 views)

Permalink

Did you set the lacp_rate option?

e.g: bonding mode=4 miimon=100 lacp_rate=1

The default is 0 or "slow" (30 sec). set to 1 (fast, 1 sec) and see if that
makes any difference.

On Thu, Mar 27, 2014 at 11:27 AM, Rupesh Basnet <brupesh@subisu.net.np>wrote:

> Hi George,
>
> First I tried mode 4(802.3ad) at first but LACP didn't formed at all.
>
> Regards,
> --------
> Rupesh Basnet
>
>
> On 03/28/2014 12:09 AM, George B. wrote:
>
> "I have configured my linux box on mode 0"
>
> Wouldn't you need to use mode 4?
>
> 802.3ad or 4
>
> IEEE 802.3ad Dynamic link aggregation. Creates
> aggregation groups that share the same speed and
> duplex settings. Utilizes all slaves in the active
> aggregator according to the 802.3ad specification.
>
>
>
> On Thu, Mar 27, 2014 at 10:44 AM, Rupesh Basnet <brupesh@subisu.net.np>wrote:
>
>> Hi All,
>>
>> I have been trying LACP between Foundry and Linux machine since some days
>> but not able to figure out the issue and resolve the issue. I didn't see
>> much configurations thats needs to be done on foundry. I have configured my
>> linux box on mode 0 and bonding seems to be up but still I am not able to
>> pass traffic through it.
>>
>> Serveriron Configurations:
>>
>> interface ethernet 4/2
>> port-name **LACP-with-Linux***
>> no spanning-tree
>> link-aggregate active
>> !
>> interface ethernet 4/3
>> port-name **LACP-with-Linux***
>> no spanning-tree
>> link-aggregate active
>> !
>>
>>
>> SI#show version
>> SW: Version 09.3.00eTD4 Copyright (c) 1996-2003 Foundry Networks, Inc.
>> Compiled on Jun 28 2005 at 18:57:32 labeled as WXR09300e
>> HW: ServerIronGT E-2 Router, SYSIF version 21, Serial #: Non-exist
>> ==========================================================================
>> SL 1: B0GMR WSM2 Management Module, SYSIF 2, M6, ACTIVE
>> Serial #: CH39040033
>> 0 MB SHM, 2 Application Processors
>> 16384 KB BRAM, SMC version 1, BM version 21
>> SW: (1)09.3.00eTF2 (2)09.3.00eTF2
>> ==========================================================================
>> SL 4: J-BxGC16 JetCore Gig Copper Module, SYSIF 2
>> Serial #: CH33060779
>> 4096 KB BRAM, JetCore ASIC IGC version 49, BIA version 8a
>> 32768 KB PRAM and 2M-Bit*1 CAM for IGC 12, version 0449
>> 32768 KB PRAM and 2M-Bit*1 CAM for IGC 13, version 0449
>> 32768 KB PRAM and 2M-Bit*1 CAM for IGC 14, version 0449
>> 32768 KB PRAM and 2M-Bit*1 CAM for IGC 15, version 0449
>>
>>
>> SI#show link-aggregate
>> System ID: 000c.db5d.b066
>> Long timeout: 90, default: 90
>> Short timeout: 3, default: 3
>> Port [Sys P] [Port P] [ Key ]
>> [Act][Tio][Agg][Syn][Col][Dis][Def][Exp][Ope]
>> 4/2 1 1 483 Yes S Agg Syn Col Dis Def No Ope
>> 4/3 1 1 483 Yes S Agg Syn Col Dis Def No Ope
>>
>>
>> Status of linux machine:
>> MII Status: up
>> MII Polling Interval (ms): 30000
>> Up Delay (ms): 0
>> Down Delay (ms): 0
>>
>> Slave Interface: eth0
>> MII Status: up
>> Speed: 1000 Mbps
>> Duplex: full
>> Link Failure Count: 0
>> Permanent HW addr: 00:e0:ed:2b:a1:99
>> Slave queue ID: 0
>>
>> Slave Interface: eth1
>> MII Status: up
>> Speed: 1000 Mbps
>> Duplex: full
>> Link Failure Count: 0
>> Permanent HW addr: 00:e0:ed:2b:a1:98
>> Slave queue ID: 0
>>
>> --
>> Regards,
>> --------
>> Rupesh Basnet
>>
>> _______________________________________________
>> foundry-nsp mailing list
>> foundry-nsp@puck.nether.net
>> http://puck.nether.net/mailman/listinfo/foundry-nsp
>>
>
>
>

Re: LACP with Foundry and Linux Machine [ In reply to ]

brupesh at subisu

Mar 27, 2014, 9:17 PM

Post #6 of 18 (8537 views)

Permalink

Yes, I did that as well.

Regards,
--------
Rupesh Basnet

On 03/28/2014 12:03 AM, Nick Hilliard wrote:
> On 27/03/2014 17:44, Rupesh Basnet wrote:
>> I have been trying LACP between Foundry and Linux machine since some days
>> but not able to figure out the issue and resolve the issue. I didn't see
>> much configurations thats needs to be done on foundry. I have configured my
>> linux box on mode 0 and bonding seems to be up but still I am not able to
>> pass traffic through it.
> have you assigned the correct untagged vlan to e4/2 and 4/3?
>
> Nick
>

_______________________________________________
foundry-nsp mailing list
foundry-nsp@puck.nether.net
http://puck.nether.net/mailman/listinfo/foundry-nsp

Re: LACP with Foundry and Linux Machine [ In reply to ]

brupesh at subisu

Mar 27, 2014, 9:58 PM

Post #7 of 18 (8538 views)

Permalink

Hi George,

lacp_rate was set to slow. I tried changing it to fast as well but still
couldn't get any success.

Regards,
--------
Rupesh Basnet

On 03/28/2014 12:17 AM, George B. wrote:
> Did you set the lacp_rate option?
>
> e.g: bonding mode=4miimon=100lacp_rate=1
>
> The default is 0 or "slow" (30 sec). set to 1 (fast, 1 sec) and see if
> that makes any difference.
>
>
>
> On Thu, Mar 27, 2014 at 11:27 AM, Rupesh Basnet <brupesh@subisu.net.np
> <mailto:brupesh@subisu.net.np>> wrote:
>
> Hi George,
>
> First I tried mode 4(802.3ad) at first but LACP didn't formed at all.
>
> Regards,
> --------
> Rupesh Basnet
>
> On 03/28/2014 12:09 AM, George B. wrote:
>> "I have configured my linux box on mode 0"
>>
>> Wouldn't you need to use mode 4?
>>
>> 802.3ad or 4
>>
>> IEEE 802.3ad Dynamic link aggregation. Creates
>> aggregation groups that share the same speed and
>> duplex settings. Utilizes all slaves in the active
>> aggregator according to the 802.3ad specification.
>>
>>
>> On Thu, Mar 27, 2014 at 10:44 AM, Rupesh Basnet
>> <brupesh@subisu.net.np <mailto:brupesh@subisu.net.np>> wrote:
>>
>> Hi All,
>>
>> I have been trying LACP between Foundry and Linux machine
>> since some days but not able to figure out the issue and
>> resolve the issue. I didn't see much configurations thats
>> needs to be done on foundry. I have configured my linux box
>> on mode 0 and bonding seems to be up but still I am not able
>> to pass traffic through it.
>>
>> Serveriron Configurations:
>>
>> interface ethernet 4/2
>> port-name **LACP-with-Linux***
>> no spanning-tree
>> link-aggregate active
>> !
>> interface ethernet 4/3
>> port-name **LACP-with-Linux***
>> no spanning-tree
>> link-aggregate active
>> !
>>
>>
>> SI#show version
>> SW: Version 09.3.00eTD4 Copyright (c) 1996-2003 Foundry
>> Networks, Inc.
>> Compiled on Jun 28 2005 at 18:57:32 labeled as WXR09300e
>> HW: ServerIronGT E-2 Router, SYSIF version 21, Serial #:
>> Non-exist
>> ==========================================================================
>> SL 1: B0GMR WSM2 Management Module, SYSIF 2, M6, ACTIVE
>> Serial #: CH39040033
>> 0 MB SHM, 2 Application Processors
>> 16384 KB BRAM, SMC version 1, BM version 21
>> SW: (1)09.3.00eTF2 (2)09.3.00eTF2
>> ==========================================================================
>> SL 4: J-BxGC16 JetCore Gig Copper Module, SYSIF 2
>> Serial #: CH33060779
>> 4096 KB BRAM, JetCore ASIC IGC version 49, BIA version 8a
>> 32768 KB PRAM and 2M-Bit*1 CAM for IGC 12, version 0449
>> 32768 KB PRAM and 2M-Bit*1 CAM for IGC 13, version 0449
>> 32768 KB PRAM and 2M-Bit*1 CAM for IGC 14, version 0449
>> 32768 KB PRAM and 2M-Bit*1 CAM for IGC 15, version 0449
>>
>>
>> SI#show link-aggregate
>> System ID: 000c.db5d.b066
>> Long timeout: 90, default: 90
>> Short timeout: 3, default: 3
>> Port [Sys P] [Port P] [ Key ]
>> [Act][Tio][Agg][Syn][Col][Dis][Def][Exp][Ope]
>> 4/2 1 1 483 Yes S Agg Syn Col Dis
>> Def No Ope
>> 4/3 1 1 483 Yes S Agg Syn Col Dis
>> Def No Ope
>>
>>
>> Status of linux machine:
>> MII Status: up
>> MII Polling Interval (ms): 30000
>> Up Delay (ms): 0
>> Down Delay (ms): 0
>>
>> Slave Interface: eth0
>> MII Status: up
>> Speed: 1000 Mbps
>> Duplex: full
>> Link Failure Count: 0
>> Permanent HW addr: 00:e0:ed:2b:a1:99
>> Slave queue ID: 0
>>
>> Slave Interface: eth1
>> MII Status: up
>> Speed: 1000 Mbps
>> Duplex: full
>> Link Failure Count: 0
>> Permanent HW addr: 00:e0:ed:2b:a1:98
>> Slave queue ID: 0
>>
>> --
>> Regards,
>> --------
>> Rupesh Basnet
>>
>> _______________________________________________
>> foundry-nsp mailing list
>> foundry-nsp@puck.nether.net <mailto:foundry-nsp@puck.nether.net>
>> http://puck.nether.net/mailman/listinfo/foundry-nsp
>>
>>
>
>

Re: LACP with Foundry and Linux Machine [ In reply to ]

erich at hoh

Mar 27, 2014, 11:39 PM

Post #8 of 18 (8528 views)

Permalink

Hi Rupesh,

I guess you hit the hardware limits when you try to build the primary
port 4/2 on this "old" foundry device.

<snip>
Each trunk group must start with a primary port. The primary port is
always the lowest number in the port range. For example, trunk groups on
a Chassis device can have the following primary ports:
2-port trunk groups: 1, 3, 5, 7, 9, 11,13, 15, 17, 19, 21, 23
</snip>

Also you can choose the the trunk type switch (default) or server which
influence the hash algorithm for the balancing. See below:

http://www.foundrynet.com/services/documentation/sribcg/current/Trunking.html

Regards

Erich

Am 27.03.14 18:44, schrieb Rupesh Basnet:
> Hi All,
>
> I have been trying LACP between Foundry and Linux machine since some
> days but not able to figure out the issue and resolve the issue. I
> didn't see much configurations thats needs to be done on foundry. I have
> configured my linux box on mode 0 and bonding seems to be up but still I
> am not able to pass traffic through it.
>
> Serveriron Configurations:
>
> interface ethernet 4/2
> port-name **LACP-with-Linux***
> no spanning-tree
> link-aggregate active
> !
> interface ethernet 4/3
> port-name **LACP-with-Linux***
> no spanning-tree
> link-aggregate active
> !
>
>
> SI#show version
> SW: Version 09.3.00eTD4 Copyright (c) 1996-2003 Foundry Networks, Inc.
> Compiled on Jun 28 2005 at 18:57:32 labeled as WXR09300e
> HW: ServerIronGT E-2 Router, SYSIF version 21, Serial #: Non-exist
> ==========================================================================
> SL 1: B0GMR WSM2 Management Module, SYSIF 2, M6, ACTIVE
> Serial #: CH39040033
> 0 MB SHM, 2 Application Processors
> 16384 KB BRAM, SMC version 1, BM version 21
> SW: (1)09.3.00eTF2 (2)09.3.00eTF2
> ==========================================================================
> SL 4: J-BxGC16 JetCore Gig Copper Module, SYSIF 2
> Serial #: CH33060779
> 4096 KB BRAM, JetCore ASIC IGC version 49, BIA version 8a
> 32768 KB PRAM and 2M-Bit*1 CAM for IGC 12, version 0449
> 32768 KB PRAM and 2M-Bit*1 CAM for IGC 13, version 0449
> 32768 KB PRAM and 2M-Bit*1 CAM for IGC 14, version 0449
> 32768 KB PRAM and 2M-Bit*1 CAM for IGC 15, version 0449
>
>
> SI#show link-aggregate
> System ID: 000c.db5d.b066
> Long timeout: 90, default: 90
> Short timeout: 3, default: 3
> Port [Sys P] [Port P] [ Key ]
> [Act][Tio][Agg][Syn][Col][Dis][Def][Exp][Ope]
> 4/2 1 1 483 Yes S Agg Syn Col Dis Def No Ope
> 4/3 1 1 483 Yes S Agg Syn Col Dis Def No Ope
>
>
> Status of linux machine:
> MII Status: up
> MII Polling Interval (ms): 30000
> Up Delay (ms): 0
> Down Delay (ms): 0
>
> Slave Interface: eth0
> MII Status: up
> Speed: 1000 Mbps
> Duplex: full
> Link Failure Count: 0
> Permanent HW addr: 00:e0:ed:2b:a1:99
> Slave queue ID: 0
>
> Slave Interface: eth1
> MII Status: up
> Speed: 1000 Mbps
> Duplex: full
> Link Failure Count: 0
> Permanent HW addr: 00:e0:ed:2b:a1:98
> Slave queue ID: 0
>
_______________________________________________
foundry-nsp mailing list
foundry-nsp@puck.nether.net
http://puck.nether.net/mailman/listinfo/foundry-nsp

Re: LACP with Foundry and Linux Machine [ In reply to ]

brupesh at subisu

Mar 28, 2014, 3:54 AM

Post #9 of 18 (8526 views)

Permalink

Hi Erich,

I am to trunk two interfaces 4/3 and 4/4 but on the trunk status still
the operational trunk capacity is 1G. Can you please help me further? I
could not get much information on the link you shared for the switch
type trunk for load balancing.

SI#show trunk

Configured trunks:

Trunk ID: 98
Type: Switch
Ports_Configured: 2
Primary Port Monitored: Jointly

Ports 4/3 4/4
Port Names none none
Port_Status enable enable
Monitor off off
Mirror Port N/A N/A
Monitor Dir N/A N/A

Operational trunks:

Trunk ID: 98
Type: Switch
Duplex: Full
Speed: 1G
Tag: No
Priority: level0
Active Ports: 2

Ports 4/3 4/4
Link_Status active active
Load Sharing
Mac Address 1 0
IP 0 0
IPX 0 0
Apple Talk 0 0
Multicast 0 0

Regards,
--------
Rupesh Basnet

On 03/28/2014 12:24 PM, Erich Hohermuth wrote:
> Hi Rupesh,
>
> I guess you hit the hardware limits when you try to build the primary
> port 4/2 on this "old" foundry device.
>
> <snip>
> Each trunk group must start with a primary port. The primary port is
> always the lowest number in the port range. For example, trunk groups on
> a Chassis device can have the following primary ports:
> 2-port trunk groups: 1, 3, 5, 7, 9, 11,13, 15, 17, 19, 21, 23
> </snip>
>
> Also you can choose the the trunk type switch (default) or server which
> influence the hash algorithm for the balancing. See below:
>
> http://www.foundrynet.com/services/documentation/sribcg/current/Trunking.html
>
> Regards
>
> Erich
>
> Am 27.03.14 18:44, schrieb Rupesh Basnet:
>> Hi All,
>>
>> I have been trying LACP between Foundry and Linux machine since some
>> days but not able to figure out the issue and resolve the issue. I
>> didn't see much configurations thats needs to be done on foundry. I have
>> configured my linux box on mode 0 and bonding seems to be up but still I
>> am not able to pass traffic through it.
>>
>> Serveriron Configurations:
>>
>> interface ethernet 4/2
>> port-name **LACP-with-Linux***
>> no spanning-tree
>> link-aggregate active
>> !
>> interface ethernet 4/3
>> port-name **LACP-with-Linux***
>> no spanning-tree
>> link-aggregate active
>> !
>>
>>
>> SI#show version
>> SW: Version 09.3.00eTD4 Copyright (c) 1996-2003 Foundry Networks, Inc.
>> Compiled on Jun 28 2005 at 18:57:32 labeled as WXR09300e
>> HW: ServerIronGT E-2 Router, SYSIF version 21, Serial #: Non-exist
>> ==========================================================================
>> SL 1: B0GMR WSM2 Management Module, SYSIF 2, M6, ACTIVE
>> Serial #: CH39040033
>> 0 MB SHM, 2 Application Processors
>> 16384 KB BRAM, SMC version 1, BM version 21
>> SW: (1)09.3.00eTF2 (2)09.3.00eTF2
>> ==========================================================================
>> SL 4: J-BxGC16 JetCore Gig Copper Module, SYSIF 2
>> Serial #: CH33060779
>> 4096 KB BRAM, JetCore ASIC IGC version 49, BIA version 8a
>> 32768 KB PRAM and 2M-Bit*1 CAM for IGC 12, version 0449
>> 32768 KB PRAM and 2M-Bit*1 CAM for IGC 13, version 0449
>> 32768 KB PRAM and 2M-Bit*1 CAM for IGC 14, version 0449
>> 32768 KB PRAM and 2M-Bit*1 CAM for IGC 15, version 0449
>>
>>
>> SI#show link-aggregate
>> System ID: 000c.db5d.b066
>> Long timeout: 90, default: 90
>> Short timeout: 3, default: 3
>> Port [Sys P] [Port P] [ Key ]
>> [Act][Tio][Agg][Syn][Col][Dis][Def][Exp][Ope]
>> 4/2 1 1 483 Yes S Agg Syn Col Dis Def No Ope
>> 4/3 1 1 483 Yes S Agg Syn Col Dis Def No Ope
>>
>>
>> Status of linux machine:
>> MII Status: up
>> MII Polling Interval (ms): 30000
>> Up Delay (ms): 0
>> Down Delay (ms): 0
>>
>> Slave Interface: eth0
>> MII Status: up
>> Speed: 1000 Mbps
>> Duplex: full
>> Link Failure Count: 0
>> Permanent HW addr: 00:e0:ed:2b:a1:99
>> Slave queue ID: 0
>>
>> Slave Interface: eth1
>> MII Status: up
>> Speed: 1000 Mbps
>> Duplex: full
>> Link Failure Count: 0
>> Permanent HW addr: 00:e0:ed:2b:a1:98
>> Slave queue ID: 0
>>
> _______________________________________________
> foundry-nsp mailing list
> foundry-nsp@puck.nether.net
> http://puck.nether.net/mailman/listinfo/foundry-nsp

_______________________________________________
foundry-nsp mailing list
foundry-nsp@puck.nether.net
http://puck.nether.net/mailman/listinfo/foundry-nsp

Re: LACP with Foundry and Linux Machine [ In reply to ]

erich at hoh

Mar 28, 2014, 11:00 PM

Post #10 of 18 (8534 views)

Permalink

Hi Rupesh,

According to your post which is only one side I can assume the
following.

> Trunk ID: 98
> Type: Switch
> Duplex: Full
> Speed: 1G
> Tag: No
> Priority: level0
> Active Ports: 2

This looks fine to me (Speed = Port Speed)

> Ports 4/3 4/4
> Link_Status active active

LACP Information is missing. What is your linux doing now (cat
/proc/net/bonding/bond0) ?

> Load Sharing
> Mac Address 1 0

Only one mac is visible. I guess you should switch to mode 4 and adjust
the lacp timers like George said.
Be careful and read your man pages carefully, cause linux distros are
likely to change their syntax after a dist upgrade.
Please bare in mind the limits of ethernet interface bonding in general
(unknown, multicast and unicast hashing).

Regards
Erich
_______________________________________________
foundry-nsp mailing list
foundry-nsp@puck.nether.net
http://puck.nether.net/mailman/listinfo/foundry-nsp

Re: LACP with Foundry and Linux Machine [ In reply to ]

brupesh at subisu

Mar 29, 2014, 10:12 PM

Post #11 of 18 (8528 views)

Permalink

Hi Erich,

I tried changing LACP rates with mode 4 but status was the same and
linux box was also not able to get partner MAC on its aggregation
status. I tried with balance-rr mode as well but same, only single MAC
was obtained under Load Balancing on foundry. I can pass traffic from
the bonded interface but still don't have any idea if those interfaces
are really bonded and load balance is happening.

*Status on Linux Box with Mode1:*
Bonding Mode: load balancing (round-robin)
MII Status: up
MII Polling Interval (ms): 30000
Up Delay (ms): 0
Down Delay (ms): 0

Slave Interface: eth0
MII Status: up
Speed: 1000 Mbps
Duplex: full
Link Failure Count: 0
Permanent HW addr: 00:e0:ed:2b:a1:96
Slave queue ID: 0

Slave Interface: eth1
MII Status: up
Speed: 1000 Mbps
Duplex: full
Link Failure Count: 0
Permanent HW addr: 00:e0:ed:2b:a1:97
Slave queue ID: 0

*Status on Linux Box with mode 4:*
Bonding Mode: IEEE 802.3ad Dynamic link aggregation
Transmit Hash Policy: layer2 (0)
MII Status: up
MII Polling Interval (ms): 30000
Up Delay (ms): 0
Down Delay (ms): 0

802.3ad info
LACP rate: slow
Min links: 0
Aggregator selection policy (ad_select): stable
Active Aggregator Info:
Aggregator ID: 1
Number of ports: 1
Actor Key: 17
Partner Key: 1
Partner Mac Address: 00:00:00:00:00:00

Slave Interface: eth0
MII Status: up
Speed: 1000 Mbps
Duplex: full
Link Failure Count: 0
Permanent HW addr: 00:e0:ed:2b:a1:96
Aggregator ID: 1
Slave queue ID: 0

Slave Interface: eth1
MII Status: up
Speed: 1000 Mbps
Duplex: full
Link Failure Count: 0
Permanent HW addr: 00:e0:ed:2b:a1:97
Aggregator ID: 2
Slave queue ID: 0

Regards,
--------
Rupesh Basnet

System Operations
Subisu Cablenet (P.) Ltd.
148 Thirbum Sadak
Baluwatar, Kathmandu
Nepal

T: 00977 1 4429616/17 Ext.: 322,323(4412832-Direct Line)
F: 00977 1 4430572

http://www.subisu.net.np

(An ISO 9001:2008 Certified Company)

On 03/29/2014 11:45 AM, Erich Hohermuth wrote:
> Hi Rupesh,
>
> According to your post which is only one side I can assume the
> following.
>
>> Trunk ID: 98
>> Type: Switch
>> Duplex: Full
>> Speed: 1G
>> Tag: No
>> Priority: level0
>> Active Ports: 2
> This looks fine to me (Speed = Port Speed)
>
>
>> Ports 4/3 4/4
>> Link_Status active active
> LACP Information is missing. What is your linux doing now (cat
> /proc/net/bonding/bond0) ?
>
>> Load Sharing
>> Mac Address 1 0
> Only one mac is visible. I guess you should switch to mode 4 and adjust
> the lacp timers like George said.
> Be careful and read your man pages carefully, cause linux distros are
> likely to change their syntax after a dist upgrade.
> Please bare in mind the limits of ethernet interface bonding in general
> (unknown, multicast and unicast hashing).
>
> Regards
> Erich

Re: LACP with Foundry and Linux Machine [ In reply to ]

raoul at bhatia

Mar 29, 2014, 10:56 PM

Post #12 of 18 (8512 views)

Permalink

On 30 March 2014 07:12:08 CEST, Rupesh Basnet <brupesh@subisu.net.np> wrote:
>Hi Erich,
>
>I tried changing LACP rates with mode 4 but status was the same and
>linux box was also not able to get partner MAC on its aggregation
>status. I tried with balance-rr mode as well but same, only single MAC
>was obtained under Load Balancing on foundry. I can pass traffic from
>the bonded interface but still don't have any idea if those interfaces
>are really bonded and load balance is happening.
Hi,

It has been some time since I was last configuring bonding but I *think* that balancing happened on the basis of MAC addresses, (or even ports?) or some other part of the packet header.

So while testing, you might see balancing only after using multiple streams and source/destination MACs.

Cheers,
Raoul
--
Raoul Bhatia
_______________________________________________
foundry-nsp mailing list
foundry-nsp@puck.nether.net
http://puck.nether.net/mailman/listinfo/foundry-nsp

Re: LACP with Foundry and Linux Machine [ In reply to ]

brupesh at subisu

Mar 31, 2014, 9:56 PM

Post #13 of 18 (8520 views)

Permalink

Hi Raoul,

With trunk on foundry and mode 4 on linux box with load balancing on l2
hash mode, load balance was happening is asymmetric way. IN traffic was
seen from one and OUT traffic was seen from another interface. I think
server mode on foundry might help but it seems we need to reload to
reconfigure it. Are there any way out for proper load balancing?

Regards,
--------
Rupesh Basnet

System Operations
Subisu Cablenet (P.) Ltd.
148 Thirbum Sadak
Baluwatar, Kathmandu
Nepal

T: 00977 1 4429616/17 Ext.: 322,323(4412832-Direct Line)
F: 00977 1 4430572

http://www.subisu.net.np

(An ISO 9001:2008 Certified Company)

On 03/30/2014 11:41 AM, Raoul Bhatia wrote:
> On 30 March 2014 07:12:08 CEST, Rupesh Basnet <brupesh@subisu.net.np> wrote:
>> Hi Erich,
>>
>> I tried changing LACP rates with mode 4 but status was the same and
>> linux box was also not able to get partner MAC on its aggregation
>> status. I tried with balance-rr mode as well but same, only single MAC
>> was obtained under Load Balancing on foundry. I can pass traffic from
>> the bonded interface but still don't have any idea if those interfaces
>> are really bonded and load balance is happening.
> Hi,
>
> It has been some time since I was last configuring bonding but I *think* that balancing happened on the basis of MAC addresses, (or even ports?) or some other part of the packet header.
>
> So while testing, you might see balancing only after using multiple streams and source/destination MACs.
>
> Cheers,
> Raoul

_______________________________________________
foundry-nsp mailing list
foundry-nsp@puck.nether.net
http://puck.nether.net/mailman/listinfo/foundry-nsp

Re: LACP with Foundry and Linux Machine [ In reply to ]

frnkblk at iname

Mar 31, 2014, 10:08 PM

Post #14 of 18 (8514 views)

Permalink

You can't control the hashing on inbound, but you can configure the
outbound.

For the Linux machine there appears to be two options: layer2 or layer3+4
(http://www.cyberciti.biz/howto/question/static/linux-ethernet-bonding-drive
r-howto.php).

Frank

-----Original Message-----
From: foundry-nsp [mailto:foundry-nsp-bounces@puck.nether.net] On Behalf Of
Rupesh Basnet
Sent: Monday, March 31, 2014 11:57 PM
To: Raoul Bhatia; Erich Hohermuth
Cc: foundry-nsp@puck.nether.net
Subject: Re: [f-nsp] LACP with Foundry and Linux Machine

Hi Raoul,

With trunk on foundry and mode 4 on linux box with load balancing on l2
hash mode, load balance was happening is asymmetric way. IN traffic was
seen from one and OUT traffic was seen from another interface. I think
server mode on foundry might help but it seems we need to reload to
reconfigure it. Are there any way out for proper load balancing?

Regards,
--------
Rupesh Basnet

System Operations
Subisu Cablenet (P.) Ltd.
148 Thirbum Sadak
Baluwatar, Kathmandu
Nepal

T: 00977 1 4429616/17 Ext.: 322,323(4412832-Direct Line)
F: 00977 1 4430572

http://www.subisu.net.np

(An ISO 9001:2008 Certified Company)

On 03/30/2014 11:41 AM, Raoul Bhatia wrote:
> On 30 March 2014 07:12:08 CEST, Rupesh Basnet <brupesh@subisu.net.np>
wrote:
>> Hi Erich,
>>
>> I tried changing LACP rates with mode 4 but status was the same and
>> linux box was also not able to get partner MAC on its aggregation
>> status. I tried with balance-rr mode as well but same, only single MAC
>> was obtained under Load Balancing on foundry. I can pass traffic from
>> the bonded interface but still don't have any idea if those interfaces
>> are really bonded and load balance is happening.
> Hi,
>
> It has been some time since I was last configuring bonding but I *think*
that balancing happened on the basis of MAC addresses, (or even ports?) or
some other part of the packet header.
>
> So while testing, you might see balancing only after using multiple
streams and source/destination MACs.
>
> Cheers,
> Raoul

_______________________________________________
foundry-nsp mailing list
foundry-nsp@puck.nether.net
http://puck.nether.net/mailman/listinfo/foundry-nsp

_______________________________________________
foundry-nsp mailing list
foundry-nsp@puck.nether.net
http://puck.nether.net/mailman/listinfo/foundry-nsp

Re: LACP with Foundry and Linux Machine [ In reply to ]

georgeb at gmail

Mar 31, 2014, 10:13 PM

Post #15 of 18 (8516 views)

Permalink

I tend to use layer3+4 because for the average server the majority of the
traffic is going to its default gateway so that always gets hashed to the
same link when using layer2 hashing. Layer3-4 will at least hash it based
on destination source port and IP address and as most client connections
come from a wide range of source ports, that more effectively balances the
traffic across the links.

On Mon, Mar 31, 2014 at 10:08 PM, Frank Bulk <frnkblk@iname.com> wrote:

> You can't control the hashing on inbound, but you can configure the
> outbound.
>
> For the Linux machine there appears to be two options: layer2 or layer3+4
> (
> http://www.cyberciti.biz/howto/question/static/linux-ethernet-bonding-drive
> r-howto.php).
>
> Frank
>
> -----Original Message-----
> From: foundry-nsp [mailto:foundry-nsp-bounces@puck.nether.net] On Behalf
> Of
> Rupesh Basnet
> Sent: Monday, March 31, 2014 11:57 PM
> To: Raoul Bhatia; Erich Hohermuth
> Cc: foundry-nsp@puck.nether.net
> Subject: Re: [f-nsp] LACP with Foundry and Linux Machine
>
> Hi Raoul,
>
> With trunk on foundry and mode 4 on linux box with load balancing on l2
> hash mode, load balance was happening is asymmetric way. IN traffic was
> seen from one and OUT traffic was seen from another interface. I think
> server mode on foundry might help but it seems we need to reload to
> reconfigure it. Are there any way out for proper load balancing?
>
> Regards,
> --------
> Rupesh Basnet
>
> System Operations
> Subisu Cablenet (P.) Ltd.
> 148 Thirbum Sadak
> Baluwatar, Kathmandu
> Nepal
>
>
> T: 00977 1 4429616/17 Ext.: 322,323(4412832-Direct Line)
> F: 00977 1 4430572
>
> http://www.subisu.net.np
>
> (An ISO 9001:2008 Certified Company)
>
> On 03/30/2014 11:41 AM, Raoul Bhatia wrote:
> > On 30 March 2014 07:12:08 CEST, Rupesh Basnet <brupesh@subisu.net.np>
> wrote:
> >> Hi Erich,
> >>
> >> I tried changing LACP rates with mode 4 but status was the same and
> >> linux box was also not able to get partner MAC on its aggregation
> >> status. I tried with balance-rr mode as well but same, only single MAC
> >> was obtained under Load Balancing on foundry. I can pass traffic from
> >> the bonded interface but still don't have any idea if those interfaces
> >> are really bonded and load balance is happening.
> > Hi,
> >
> > It has been some time since I was last configuring bonding but I *think*
> that balancing happened on the basis of MAC addresses, (or even ports?) or
> some other part of the packet header.
> >
> > So while testing, you might see balancing only after using multiple
> streams and source/destination MACs.
> >
> > Cheers,
> > Raoul
>
> _______________________________________________
> foundry-nsp mailing list
> foundry-nsp@puck.nether.net
> http://puck.nether.net/mailman/listinfo/foundry-nsp
>
>
> _______________________________________________
> foundry-nsp mailing list
> foundry-nsp@puck.nether.net
> http://puck.nether.net/mailman/listinfo/foundry-nsp
>

Re: LACP with Foundry and Linux Machine [ In reply to ]

brupesh at subisu

Apr 4, 2014, 6:50 AM

Post #16 of 18 (8497 views)

Permalink

Hi All,

L3 hashing mode is not helping on my case, there are only one way
traffic on two interfaces and I am not able to figure out issue. Is
server mode trunk a L3 hashing mode on foundry?

Regards,
--------
Rupesh Basnet

On 2014-04-01 10:58, George B. wrote:
> I tend to use layer3+4 because for the average server the majority of
> the traffic is going to its default gateway so that always gets
> hashed
> to the same link when using layer2 hashing. Â Layer3-4 will at least
> hash it based on destination source port and IP address and as most
> client connections come from a wide range of source ports, that more
> effectively balances the traffic across the links.
>
> On Mon, Mar 31, 2014 at 10:08 PM, Frank Bulk <frnkblk@iname.com [9]>
> wrote:
>
>> You can't control the hashing on inbound, but you can configure the
>> outbound.
>>
>> For the Linux machine there appears to be two options: layer2 or
>> layer3+4
>>
>
> (http://www.cyberciti.biz/howto/question/static/linux-ethernet-bonding-drive
>> r-howto.php).
>>
>> Frank
>>
>> -----Original Message-----
>> From: foundry-nsp [mailto:foundry-nsp-bounces@puck.nether.net [1]]
>> On Behalf Of
>> Rupesh Basnet
>> Sent: Monday, March 31, 2014 11:57 PM
>> To: Raoul Bhatia; Erich Hohermuth
>> Cc: foundry-nsp@puck.nether.net [2]
>> Subject: Re: [f-nsp] LACP with Foundry and Linux Machine
>>
>> Hi Raoul,
>>
>> With trunk on foundry and mode 4 on linux box with load balancing
>> on l2
>> hash mode, load balance was happening is asymmetric way. IN traffic
>> was
>> seen from one and Â OUT traffic was seen from another interface. I
>> think
>> server mode on foundry might help but it seems we need to reload to
>> reconfigure it. Are there any way out for proper load balancing?
>>
>> Regards,
>> --------
>> Rupesh Basnet
>>
>> System Operations
>> Subisu Cablenet (P.) Ltd.
>> 148 Thirbum Sadak
>> Baluwatar, Kathmandu
>> Nepal
>>
>> T: 00977 1 4429616/17 Ext.: 322,323(4412832-Direct Line)
>> F: 00977 1 4430572
>>
>> http://www.subisu.net.np [3]
>>
>> (An ISO 9001:2008 Certified Company)
>>
>> On 03/30/2014 11:41 AM, Raoul Bhatia wrote:
>> > On 30 March 2014 07:12:08 CEST, Rupesh Basnet
>> <brupesh@subisu.net.np [4]>
>> wrote:
>> >> Hi Erich,
>> >>
>> >> I tried changing LACP rates with mode 4 but status was the same
>> and
>> >> linux box was also not able to get partner MAC on its
>> aggregation
>> >> status. I tried with balance-rr mode as well but same, only
>> single MAC
>> >> was obtained under Load Balancing on foundry. I can pass traffic
>> from
>> >> the bonded interface but still don't have any idea if those
>> interfaces
>> >> are really bonded and load balance is happening.
>> > Hi,
>> >
>> > It has been some time since I was last configuring bonding but I
>> *think*
>> that balancing happened on the basis of MAC addresses, (or even
>> ports?) or
>> some other part of the packet header.
>> >
>> > So while testing, you might see balancing only after using
>> multiple
>> streams and source/destination MACs.
>> >
>> > Cheers,
>> > Raoul
>>
>> _______________________________________________
>> foundry-nsp mailing list
>> foundry-nsp@puck.nether.net [5]
>> http://puck.nether.net/mailman/listinfo/foundry-nsp [6]
>>
>> _______________________________________________
>> foundry-nsp mailing list
>> foundry-nsp@puck.nether.net [7]
>> http://puck.nether.net/mailman/listinfo/foundry-nsp [8]
>
>
>
> Links:
> ------
> [1] mailto:foundry-nsp-bounces@puck.nether.net
> [2] mailto:foundry-nsp@puck.nether.net
> [3] http://www.subisu.net.np
> [4] mailto:brupesh@subisu.net.np
> [5] mailto:foundry-nsp@puck.nether.net
> [6] http://puck.nether.net/mailman/listinfo/foundry-nsp
> [7] mailto:foundry-nsp@puck.nether.net
> [8] http://puck.nether.net/mailman/listinfo/foundry-nsp
> [9] mailto:frnkblk@iname.com

_______________________________________________
foundry-nsp mailing list
foundry-nsp@puck.nether.net
http://puck.nether.net/mailman/listinfo/foundry-nsp

Re: LACP with Foundry and Linux Machine [ In reply to ]

esk-puck.nether.net at esk

Apr 4, 2014, 2:05 PM

Post #17 of 18 (8517 views)

Permalink

Is this a single TCP/UDP connection? The spec requires LACP to use a
hash that guarantees a given flow will always hash to the same
interface. For LACP to work effectively, you need multiple flows.

--
Eldon Koyle
Information Technology
Utah State University
--
BOFH excuse #175:
OS swapped to disk

On Apr 04 19:35+0545, Rupesh Basnet wrote:
> Hi All,
>
> L3 hashing mode is not helping on my case, there are only one way
> traffic on two interfaces and I am not able to figure out issue. Is
> server mode trunk a L3 hashing mode on foundry?
>
>
> Regards,
> --------
> Rupesh Basnet
>
>
> On 2014-04-01 10:58, George B. wrote:
> >I tend to use layer3+4 because for the average server the majority of
> >the traffic is going to its default gateway so that always gets
> >hashed
> >to the same link when using layer2 hashing. Layer3-4 will at least
> >hash it based on destination source port and IP address and as most
> >client connections come from a wide range of source ports, that more
> >effectively balances the traffic across the links.
> >
> >On Mon, Mar 31, 2014 at 10:08 PM, Frank Bulk <frnkblk@iname.com [9]>
> >wrote:
> >
> >>You can't control the hashing on inbound, but you can configure the
> >>outbound.
> >>
> >>For the Linux machine there appears to be two options: layer2 or
> >>layer3+4
> >>
> >
> >(http://www.cyberciti.biz/howto/question/static/linux-ethernet-bonding-drive
> >>r-howto.php).
> >>
> >>Frank
> >>
> >>-----Original Message-----
> >>From: foundry-nsp [mailto:foundry-nsp-bounces@puck.nether.net [1]]
> >>On Behalf Of
> >>Rupesh Basnet
> >>Sent: Monday, March 31, 2014 11:57 PM
> >>To: Raoul Bhatia; Erich Hohermuth
> >>Cc: foundry-nsp@puck.nether.net [2]
> >>Subject: Re: [f-nsp] LACP with Foundry and Linux Machine
> >>
> >>Hi Raoul,
> >>
> >>With trunk on foundry and mode 4 on linux box with load balancing
> >>on l2
> >>hash mode, load balance was happening is asymmetric way. IN traffic
> >>was
> >>seen from one and OUT traffic was seen from another interface. I
> >>think
> >>server mode on foundry might help but it seems we need to reload to
> >>reconfigure it. Are there any way out for proper load balancing?
> >>
> >>Regards,
> >>--------
> >>Rupesh Basnet
> >>
> >>System Operations
> >>Subisu Cablenet (P.) Ltd.
> >>148 Thirbum Sadak
> >>Baluwatar, Kathmandu
> >>Nepal
> >>
> >>T: 00977 1 4429616/17 Ext.: 322,323(4412832-Direct Line)
> >>F: 00977 1 4430572
> >>
> >>http://www.subisu.net.np [3]
> >>
> >>(An ISO 9001:2008 Certified Company)
> >>
> >>On 03/30/2014 11:41 AM, Raoul Bhatia wrote:
> >>> On 30 March 2014 07:12:08 CEST, Rupesh Basnet
> >><brupesh@subisu.net.np [4]>
> >>wrote:
> >>>> Hi Erich,
> >>>>
> >>>> I tried changing LACP rates with mode 4 but status was the same
> >>and
> >>>> linux box was also not able to get partner MAC on its
> >>aggregation
> >>>> status. I tried with balance-rr mode as well but same, only
> >>single MAC
> >>>> was obtained under Load Balancing on foundry. I can pass traffic
> >>from
> >>>> the bonded interface but still don't have any idea if those
> >>interfaces
> >>>> are really bonded and load balance is happening.
> >>> Hi,
> >>>
> >>> It has been some time since I was last configuring bonding but I
> >>*think*
> >>that balancing happened on the basis of MAC addresses, (or even
> >>ports?) or
> >>some other part of the packet header.
> >>>
> >>> So while testing, you might see balancing only after using
> >>multiple
> >>streams and source/destination MACs.
> >>>
> >>> Cheers,
> >>> Raoul
> >>
> >>_______________________________________________
> >>foundry-nsp mailing list
> >>foundry-nsp@puck.nether.net [5]
> >>http://puck.nether.net/mailman/listinfo/foundry-nsp [6]
> >>
> >>_______________________________________________
> >>foundry-nsp mailing list
> >>foundry-nsp@puck.nether.net [7]
> >>http://puck.nether.net/mailman/listinfo/foundry-nsp [8]
> >
> >
> >
> >Links:
> >------
> >[1] mailto:foundry-nsp-bounces@puck.nether.net
> >[2] mailto:foundry-nsp@puck.nether.net
> >[3] http://www.subisu.net.np
> >[4] mailto:brupesh@subisu.net.np
> >[5] mailto:foundry-nsp@puck.nether.net
> >[6] http://puck.nether.net/mailman/listinfo/foundry-nsp
> >[7] mailto:foundry-nsp@puck.nether.net
> >[8] http://puck.nether.net/mailman/listinfo/foundry-nsp
> >[9] mailto:frnkblk@iname.com
>
> _______________________________________________
> foundry-nsp mailing list
> foundry-nsp@puck.nether.net
> http://puck.nether.net/mailman/listinfo/foundry-nsp
_______________________________________________
foundry-nsp mailing list
foundry-nsp@puck.nether.net
http://puck.nether.net/mailman/listinfo/foundry-nsp

Re: LACP with Foundry and Linux Machine [ In reply to ]

tcsillag at interware

Apr 5, 2014, 3:49 AM

Post #18 of 18 (8499 views)

Permalink

I think Foundry can violate that part a bit with the "Server trunk"
option, as that restriction of LACP is pretty painful with something like
iSCSI, where I'd really like to use multiple Gig interfaces for one flow.

http://www.foundrynet.com/services/documentation/sribcg/current/Trunking.html#wp112750

Tamas

> Is this a single TCP/UDP connection? The spec requires LACP to use a
> hash that guarantees a given flow will always hash to the same
> interface. For LACP to work effectively, you need multiple flows.
>
> --
> Eldon Koyle
> Information Technology
> Utah State University
> --
> BOFH excuse #175:
> OS swapped to disk
>
> On Apr 04 19:35+0545, Rupesh Basnet wrote:
>> Hi All,
>>
>> L3 hashing mode is not helping on my case, there are only one way
>> traffic on two interfaces and I am not able to figure out issue. Is
>> server mode trunk a L3 hashing mode on foundry?
>>
>>
>> Regards,
>> --------
>> Rupesh Basnet
>>
>>
>> On 2014-04-01 10:58, George B. wrote:
>> >I tend to use layer3+4 because for the average server the majority of
>> >the traffic is going to its default gateway so that always gets
>> >hashed
>> >to the same link when using layer2 hashing. Layer3-4 will at least
>> >hash it based on destination source port and IP address and as most
>> >client connections come from a wide range of source ports, that more
>> >effectively balances the traffic across the links.
>> >
>> >On Mon, Mar 31, 2014 at 10:08 PM, Frank Bulk <frnkblk@iname.com [9]>
>> >wrote:
>> >
>> >>You can't control the hashing on inbound, but you can configure the
>> >>outbound.
>> >>
>> >>For the Linux machine there appears to be two options: layer2 or
>> >>layer3+4
>> >>
>> >
>> >(http://www.cyberciti.biz/howto/question/static/linux-ethernet-bonding-drive
>> >>r-howto.php).
>> >>
>> >>Frank
>> >>
>> >>-----Original Message-----
>> >>From: foundry-nsp [mailto:foundry-nsp-bounces@puck.nether.net [1]]
>> >>On Behalf Of
>> >>Rupesh Basnet
>> >>Sent: Monday, March 31, 2014 11:57 PM
>> >>To: Raoul Bhatia; Erich Hohermuth
>> >>Cc: foundry-nsp@puck.nether.net [2]
>> >>Subject: Re: [f-nsp] LACP with Foundry and Linux Machine
>> >>
>> >>Hi Raoul,
>> >>
>> >>With trunk on foundry and mode 4 on linux box with load balancing
>> >>on l2
>> >>hash mode, load balance was happening is asymmetric way. IN traffic
>> >>was
>> >>seen from one and OUT traffic was seen from another interface. I
>> >>think
>> >>server mode on foundry might help but it seems we need to reload to
>> >>reconfigure it. Are there any way out for proper load balancing?
>> >>
>> >>Regards,
>> >>--------
>> >>Rupesh Basnet
>> >>
>> >>System Operations
>> >>Subisu Cablenet (P.) Ltd.
>> >>148 Thirbum Sadak
>> >>Baluwatar, Kathmandu
>> >>Nepal
>> >>
>> >>T: 00977 1 4429616/17 Ext.: 322,323(4412832-Direct Line)
>> >>F: 00977 1 4430572
>> >>
>> >>http://www.subisu.net.np [3]
>> >>
>> >>(An ISO 9001:2008 Certified Company)
>> >>
>> >>On 03/30/2014 11:41 AM, Raoul Bhatia wrote:
>> >>> On 30 March 2014 07:12:08 CEST, Rupesh Basnet
>> >><brupesh@subisu.net.np [4]>
>> >>wrote:
>> >>>> Hi Erich,
>> >>>>
>> >>>> I tried changing LACP rates with mode 4 but status was the same
>> >>and
>> >>>> linux box was also not able to get partner MAC on its
>> >>aggregation
>> >>>> status. I tried with balance-rr mode as well but same, only
>> >>single MAC
>> >>>> was obtained under Load Balancing on foundry. I can pass traffic
>> >>from
>> >>>> the bonded interface but still don't have any idea if those
>> >>interfaces
>> >>>> are really bonded and load balance is happening.
>> >>> Hi,
>> >>>
>> >>> It has been some time since I was last configuring bonding but I
>> >>*think*
>> >>that balancing happened on the basis of MAC addresses, (or even
>> >>ports?) or
>> >>some other part of the packet header.
>> >>>
>> >>> So while testing, you might see balancing only after using
>> >>multiple
>> >>streams and source/destination MACs.
>> >>>
>> >>> Cheers,
>> >>> Raoul
>> >>
>> >>_______________________________________________
>> >>foundry-nsp mailing list
>> >>foundry-nsp@puck.nether.net [5]
>> >>http://puck.nether.net/mailman/listinfo/foundry-nsp [6]
>> >>
>> >>_______________________________________________
>> >>foundry-nsp mailing list
>> >>foundry-nsp@puck.nether.net [7]
>> >>http://puck.nether.net/mailman/listinfo/foundry-nsp [8]
>> >
>> >
>> >
>> >Links:
>> >------
>> >[1] mailto:foundry-nsp-bounces@puck.nether.net
>> >[2] mailto:foundry-nsp@puck.nether.net
>> >[3] http://www.subisu.net.np
>> >[4] mailto:brupesh@subisu.net.np
>> >[5] mailto:foundry-nsp@puck.nether.net
>> >[6] http://puck.nether.net/mailman/listinfo/foundry-nsp
>> >[7] mailto:foundry-nsp@puck.nether.net
>> >[8] http://puck.nether.net/mailman/listinfo/foundry-nsp
>> >[9] mailto:frnkblk@iname.com
>>
>> _______________________________________________
>> foundry-nsp mailing list
>> foundry-nsp@puck.nether.net
>> http://puck.nether.net/mailman/listinfo/foundry-nsp
> _______________________________________________
> foundry-nsp mailing list
> foundry-nsp@puck.nether.net
> http://puck.nether.net/mailman/listinfo/foundry-nsp
>

_______________________________________________
foundry-nsp mailing list
foundry-nsp@puck.nether.net
http://puck.nether.net/mailman/listinfo/foundry-nsp