Re: Ping timeouts

marficadmin · ‎2008-09-11

Hi All,

We have EAS2050 and 2 interfaces, this is used for test labe and presently we are having issues the connectivity as it times out when we try to ping the box, blow is teh Network interface details, how can we check the if the network interface is fine and can we isolate the issue?

e0a: flags=4848043<UP,BROADCAST,RUNNING,MULTICAST,NOWINS> mtu 1500
inet 10.10.34.100 netmask 0xffffff00 broadcast 10.10.34.255
ether 00:a0:98:02:34:6c (100tx-fd-up) flowcontrol full
e0b: flags=4848043<UP,BROADCAST,RUNNING,MULTICAST,NOWINS> mtu 1500
inet 10.10.10.201 netmask 0xffffff00 broadcast 10.10.10.255
inet 10.10.10.200 netmask 0xffffff00 broadcast 10.10.10.255
ether 00:a0:98:02:34:6d (100tx-fd-up) flowcontrol full

The interface we are having issue is e0a.

Thanks,

vijay

kusek · ‎2008-09-11

Marfic,

I'd check for some errors on the connections

You can get that kind of information with an "ifstat -a"

Once you have that, I would also verify that the ports we're connected to are completely fine, seeing as it shows them connected at 100mb/full - I'd want to definitely verify how things are configured (auto negotiation vs manually configured, should it be gigabit, etc)

Thanks and let us know!

Christopher

marficadmin · ‎2008-09-11

Hi Chris,

Beow is the ifstat output, the connections are good physically, we have tried changing teh network cables on the interfaces as well.

-- interface e0a (17 hours, 38 minutes, 1 second) --

-- interface e0b (17 hours, 38 minutes, 1 second) --

Thanks,

Vijay

kusek · ‎2008-09-11

Vijay,

Looking at this I notice a number of CRC errors, as well as runt frames, Dribble errors, etc. - The big one being CRC errors.

First place I'd advise checking is the switch and its configuration.

If the switch itself isn't configured in a static configuration for these ports, I would advise doing that.

Along the same lines, ensure that your configuration in /etc/rc is also of a static type.

You want these two to mirror eachother.

So, lets say you have a cisco switch and this is indeed a 100mb port - Configure the port to be 100mb Full Duplex, and do the same in /etc/rc

Let me know if this helps, and if you'd like to dump the pertinent details of /etc/rc and/or of the switch config, I'd be glad to look at them!

Christopher

marficadmin · ‎2008-09-11

Hi Chris,

We tried to change the setting to Auto negosiate on both the switch and the interface e0a, the interface went down after that, we are using a GB switch on the Network;

is there any way from the command line we can change the e0a setting back to 100 MB Full?

Thanks,

Vijay

marficadmin · ‎2008-09-11

Hi Chris,

Restarting teh network service brought the interface up, but no luck with the timeouts

e0a: flags=4848043<UP,BROADCAST,RUNNING,MULTICAST,NOWINS> mtu 1500
inet 10.10.34.100 netmask 0xffffff00 broadcast 10.10.34.255
ether 00:a0:98:02:34:6c (auto-1000t-fd-up) flowcontrol full
e0b: flags=4848043<UP,BROADCAST,RUNNING,MULTICAST,NOWINS> mtu 1500
inet 10.10.10.200 netmask 0xffffff00 broadcast 10.10.10.255
inet 10.10.10.201 netmask 0xffffff00 broadcast 10.10.10.255
ether 00:a0:98:02:34:6d (100tx-fd-up) flowcontrol full

Thanks,

Vijay

kusek · ‎2008-09-12

Historically the type of errors you're seeing are typically tied to a switch configuration (misconfiguration) or negotiation issue.

So, if you test it at the physical layer as well as at the logical/configuration layer, taking it from the basics to the most complicated, you can find success.

ie: start with it at 10mb/half, 10mb/full and take it up and further, seeing at which point you start to notice failures.

If the failures persist, it may be possible to have a bad port (on either end) or an inline fault.

ifstat -a will be your friend in these circumstances (and to validate it, ifstat -z I believe can reset the details so you can keep updating)

Let us know how it goes! I'll be offline for some time now sleeping, so good luck and hope to see it working come my morning!

Thanks Vijay!

Christopher

adamfox · ‎2008-09-16

I agree with Christoper here. This is a classic case of duplex mismatch. Auto-negotiate in the 100Mb world never really worked right.

Keep in mind as well that once you get it working, ONTAP only accepts 10 pings per second. Anything more than that is considered a denial of service attack and it will stop responding for the rest of that second. This has confused some customers in the past.

marficadmin · ‎2008-09-17

Chris Thanks!!

It was an issue with the one of the port on teh switch, changing resolved the issue; thanks for the time and help 🙂 appreciate it!!

Regards,

Vijay

kusek · ‎2008-09-18

Glad to hear we could help resolve your issue!

Good luck and don't hesitate to continue to understand your filer and your operational environment Vijay!

Christopher