link aggregation - netgear gs748T

Found out about that in another thread. The GS748T does not qualify for link aggregation of OS X. There is the GS748TS model that indeed supports LACP. So now I will exchange my switch by the latter one.

Hi Jomel1,

Did you manage to get the GS748TS to work with link aggregation? I am thinking of getting the GS724TS to setup link aggregation with PowerMac G5.

This may not help but I have a Netgear GSM-7224 switch where I have been sucessful in getting ling aggregation to work just fine.

I have this working on both Intel and PPC servers.

Seems there are problems with failover too:

http://docs.info.apple.com/article.html?artnum=305066

I will try again and see with 10.4.9 if link aggregation works on a Mac Pro Intel.

curious story… have the same switch at one customer and link agg. worked fine. can't remeber, but one update (i guess 10.4.8) broke it and i couldn't get it back working. well, that server still runs on 10.4.8 and is Xserve G5 Dual 2.3 GHz.

now, which PPC server and OS version do you use on that successful setup?
was link agg. broken once (after an update)?

thx. alot for answers.

now, which PPC server and OS version do you use on that successful setup?

I am using a Dual 2.5 G5 with version 10.4.10

was link agg. broken once (after an update)?

I did need to reset the router and re-specify the Link Aggregate port.

My Server is posted below and my Switch is a DGS-3224TGR D-Link.
No pokes at the d-link please 95% of all there networking equipment that I've used at home or at offices has work flawlessly. I've had no major failures of hardware and once had to wait 1 month for a firmware update that was needed to get full functionality but the network was still working without it.
As to the issue at hand I have just implemented link aggregation again on the Pro. From all I had read and from my talks with Apple my issue was with the Mac Pro Intel and 10.4.x. I just set it up again under now 10.4.9 and all looks good on the switch web-admin, Network Status and in ifconfig. To be specific the MAC addresses in ifconfig for en0 and en1 as well as bond0 all match. Before 10.4.9 I could not get this to happen. Also the failover works now too. Before even though a bondX was created only en0 handled traffic correctly. unplug en0 and the bond would fail. Now it appears to work.
Right now there is very little traffic on the server and I will leave it configured for some stress testing. Hopefully it will hold and my users might see a bit of a performance kick. I'll keep you all posted and if you still have issues we can try to address them.

Wanted to post these so people see what it should look like now:

osxserver:~ root# netstat -i
Name Mtu Network Address Ipkts Ierrs Opkts Oerrs Coll
en0 1500 <Link#4> 00:17:f2:02:8f:06 12747 0 4942 0 0
en1 1500 <Link#5> 00:17:f2:02:8f:06 1286 0 13677 0 0
bond0 1500 <Link#7> 00:17:f2:02:8f:06 12134 0 11503 0 0
osxserver:~ root# ifconfig -a
en0: flags=8863<UP,BROADCAST,SMART,RUNNING,SIMPLEX,MULTICAST> mtu 1500
ether 00:17:f2:02:8f:06
media: autoselect (1000baseT <full-duplex>) status: active
supported media: autoselect 10baseT/UTP <half-duplex> 10baseT/UTP <full-duplex> 10baseT/UTP <full-duplex,hw-loopback> 10baseT/UTP <full-duplex,flow-control> 100baseTX <half-duplex> 100baseTX <full-duplex> 100baseTX <full-duplex,hw-loopback> 100baseTX <full-duplex,flow-control> 1000baseT <full-duplex> 1000baseT <full-duplex,hw-loopback> 1000baseT <full-duplex,flow-control>
en1: flags=8863<UP,BROADCAST,SMART,RUNNING,SIMPLEX,MULTICAST> mtu 1500
ether 00:17:f2:02:8f:06
media: autoselect (1000baseT <full-duplex>) status: active
supported media: autoselect 10baseT/UTP <half-duplex> 10baseT/UTP <full-duplex> 10baseT/UTP <full-duplex,hw-loopback> 10baseT/UTP <full-duplex,flow-control> 100baseTX <half-duplex> 100baseTX <full-duplex> 100baseTX <full-duplex,hw-loopback> 100baseTX <full-duplex,flow-control> 1000baseT <full-duplex> 1000baseT <full-duplex,hw-loopback> 1000baseT <full-duplex,flow-control>
bond0: flags=8843<UP,BROADCAST,RUNNING,SIMPLEX,MULTICAST> mtu 1500
inet6 fe80::217:f2ff:fe02:8f06%bond0 prefixlen 64 scopeid 0x7
inet 192.168.40.2 netmask 0xffffff00 broadcast 192.168.40.255
ether 00:17:f2:02:8f:06
media: autoselect (1000baseT <full-duplex>) status: active
supported media: autoselect
bond interfaces: en0 en1

In my other Post which has been archived you can see what was happening.
http://discussions.apple.com/thread.jspa?threadID=772179&tstart=0&messageID=4028 732#4028732

Well all is well with my Mac Pro & link aggregation. 10.4.9 appears to have fixed the problems some where seeing when trying to do link aggregation. I've had it up an running for 2 days now with the switch showing a good distribution of packets across both aggregated ports. What do I hear from the users. They said "Nothin but the Rain" so I told them to "get there guns and bring in the cat"
No users have reported any problem with corruption or unexpected disconnects so I'd say where good. I'd consider this case closed and if you have a link aggregation problem open a new thread.

I did need to reset the router and re-specify the
Link Aggregate port.

i guess you meant switch instead of router. my default settings for all ports is lcap-mode: enable. so i guess (as discussed in other threads), that this is kind of dynamic lcap, if the switch detects usage from a machine on the network; like OS X Server's dynamic bonding. correct, or did you specify other settings on the switch-side?

do you remember at which OS X version you discovered disfunctional bonding/lcap?

thx.

Yes, I apologize - I did mean switch.

My switch settings are identical to yours in the lcap is enabled.

I believe it was 10.4.8 that created the issue

Yes, I apologize - I did mean switch.

np, thought so. 🙂

My switch settings are identical to yours in the lcap
is enabled.

ok. thx. for your verification.

I believe it was 10.4.8 that created the issue

that would be exactly the same version. that server runs still 10.4.8 and sice then i had the issue. so i assume, i will upgrade to 10.4.10 and then re-activate link aggregation.

are there any other known issues like mySQL, PHP, LDAP on your side?

I have not had any issues though my user base is small.

The only nit that I can think of is that my external SATA drive array seems to stay spun up all the time and doesn't seem to spin down much - even with little to no load.

Ahh, the joys of a job in IT