SUMMARY: Problems with multipathing on Netra T1 AC200

Apparently, no one on the list had this problem before cause I didn't get
any responses.
Spent some time with Sun and got the solution.

Well, it turned to be a pure hardware problem, and as always, Sun has
patches to fix problems with their hardware.
My second machines, which has the same hardware and software configuration,
is still runnning w/o those patches and w/o any problems.

Apllying the following patches fixed the problem:
109882-04
110723-02
111041-02

Note that not all of these patches available to public download from
sunsolve. You need a contract.

--

Konstantin Orekhov

> -----Original Message-----
>
> Hello people!
>
> I'm having interesting problem with multipathing (MP)
> configuration on one of our Netras.
> I have 4 other machines (1 E220R, 2 Netra T1 105, 1 Netra T1
> AC200 (exact same as problem machine)) and no problem there.
>
> Here's a config:
> Netra T1 AC200;
> SunOS puka 5.8 Generic_108528-08 sun4u sparc SUNW,UltraAX-i2;
> Solaris 8 10/00 Maintenance Update 3 applied;
> Full duplex 100 Mbps is forced on eri0 and eri1 interfaces;
> eri0 connected to one switch, eri1 to another, switches are
> transparent;
>
> Here's what happens (I doing some MP testing):
> If we disable port via IOS CLI on a switches, everything is fine:
> -------------------------
> Jun 14 08:58:54 puka eri: [ID 517527 kern.info] SUNW,eri0 :
> No response from Ethernet network : Link down -- cable problem?
> Jun 14 08:59:01 puka in.mpathd[30]: [ID 594170 daemon.error]
> NIC failure detected on eri0 of group mp
> Jun 14 08:59:01 puka in.mpathd[30]: [ID 832587 daemon.error]
> Successfully failed over from NIC eri0 to NIC eri1
> Jun 14 08:59:15 puka eri: [ID 517527 kern.info] SUNW,eri0 :
> No response from Ethernet network : Link down -- cable problem?
> Jun 14 09:00:41 puka last message repeated 4 times
> Jun 14 09:00:41 puka eri: [ID 517527 kern.info] SUNW,eri0 :
> 100 Mbps full duplex link up
> Jun 14 09:01:42 puka in.mpathd[30]: [ID 299542 daemon.error]
> NIC repair detected on eri0 of group mp
> Jun 14 09:01:42 puka in.mpathd[30]: [ID 620804 daemon.error]
> Successfully failed back to NIC eri0
> --------------------------
>
> If I unplug cable from eri1 (either one - machine or switch
> side) - no prob too in.mpathd just reports that eri1 failed,
> and since eri1 is just standby interface, nothing else
> happens until I plug cable back. Once cable is in place, MP
> daemon removes eri1 from "FAILED" status.
>
> But if I remove cable from eri0 - MP reports that ALL
> interfaces went down:
> --------------------------
> Jun 13 15:03:15 puka eri: [ID 517527 kern.info] SUNW,eri0 :
> No response from Ethernet network : Link down -- cable problem?
> Jun 13 15:03:21 puka in.mpathd[28]: [ID 168056 daemon.error]
> All Interfaces in group mp have failed
> Jun 13 15:03:36 puka eri: [ID 517527 kern.info] SUNW,eri0 :
> No response from Ethernet network : Link down -- cable problem?
> ----(plug cable back)
> Jun 13 15:03:46 puka eri: [ID 517527 kern.info] SUNW,eri0 :
> 100 Mbps full duplex link up
> Jun 13 15:03:46 puka last message repeated 1 time
> Jun 13 15:04:44 puka in.mpathd[28]: [ID 237757 daemon.error]
> At least 1 interface (eri1) of group mp has repaired
> Jun 13 15:04:44 puka in.mpathd[28]: [ID 299542 daemon.error]
> NIC repair detected on eri1 of group mp
> Jun 13 15:04:47 puka in.mpathd[28]: [ID 299542 daemon.error]
> NIC repair detected on eri0 of group mp
> Jun 13 15:04:47 puka in.mpathd[28]: [ID 620804 daemon.error]
> Successfully failed back to NIC eri0
> ---------------------------
>
> Notice that eri1 is actually not losing link, therefore it's
> considered to be repaired first when plug cable back in eri0.
> So the question is does anyone have any idea why in the world
> unplugging cable from eri0 interface causing fail of all
> interfaces? Hardware problem on motherboard? In my
> understanding, breaking electrical connection causing all this crap.
>
> And one more thing. During the boot process (actually, right
> after it) all interfaces going down and up in few seconds
> (look at the time stamp):
> ---------------------------
> Jun 14 09:28:02 puka in.mpathd[30]: [ID 168056 daemon.error]
> All Interfaces in group mp have failed
> Jun 14 09:28:14 puka in.mpathd[30]: [ID 237757 daemon.error]
> At least 1 interface (eri0) of group mp has repaired
> Jun 14 09:28:14 puka in.mpathd[30]: [ID 299542 daemon.error]
> NIC repair detected on eri0 of group mp
> Jun 14 09:28:14 puka in.mpathd[30]: [ID 620804 daemon.error]
> Successfully failed back to NIC eri0
> Jun 14 09:28:15 puka in.mpathd[30]: [ID 299542 daemon.error]
> NIC repair detected on eri1 of group mp
> ---------------------------
>
> Any thoughts/suggestions/pointers will be very appreciated!
>
> TIA.
>
> --
>
> Konstantin Orekhov
>