Comware

Hi,

I have a customer with 2 HP 1810-24G V2 and 2 Hp 1810-48G switches trunked to a 2530-24G switch.

On the 18xx switches they are getting constant disconnect/reconnect of ports within a few seconds..

724 Info Mar 11 07:02:30 NIM Interface 7 is Link Down

725 Info Mar 11 07:02:34 NIM Interface 7 is Link Up

726 Info Mar 11 07:02:35 NIM Interface 7 is Link Down

727 Info Mar 11 07:02:38 NIM Interface 7 is Link Up

728 Info Mar 11 07:02:49 NIM Interface 7 is Link Down

729 Info Mar 11 07:02:52 NIM Interface 7 is Link Up

734 Info Mar 11 08:22:39 NIM Interface 18 is Link Up

735 Info Mar 11 08:22:41 NIM Interface 18 is Link Down

736 Info Mar 11 08:22:45 NIM Interface 18 is Link Up

737 Info Mar 11 08:22:57 NIM Interface 18 is Link Down

738 Info Mar 11 08:23:01 NIM Interface 18 is Link Up

739 Info Mar 11 08:26:09 NIM Interface 2 is Link Down

740 Info Mar 11 08:26:14 NIM Interface 2 is Link Up

741 Info Mar 11 08:26:15 NIM Interface 2 is Link Down

742 Info Mar 11 08:26:18 NIM Interface 2 is Link Up

743 Info Mar 11 08:26:31 NIM Interface 2 is Link Down

744 Info Mar 11 08:26:34 NIM Interface 2 is Link Up

745 Info Mar 11 08:55:32 NIM Interface 43 is Link Down

746 Info Mar 11 08:55:37 NIM Interface 43 is Link Up

747 Info Mar 11 08:55:39 NIM Interface 43 is Link Down

748 Info Mar 11 08:55:42 NIM Interface 43 is Link Up

749 Info Mar 11 08:55:55 NIM Interface 43 is Link Down

750 Info Mar 11 08:55:58 NIM Interface 43 is Link Up

751 Info Mar 11 09:01:51 NIM Interface 45 is Link Down

752 Info Mar 11 09:01:55 NIM Interface 45 is Link Up

753 Info Mar 11 09:01:58 NIM Interface 45 is Link Down

754 Info Mar 11 09:02:01 NIM Interface 45 is Link Up

755 Info Mar 11 09:02:13 NIM Interface 45 is Link Down

756 Info Mar 11 09:02:16 NIM Interface 45 is Link Up

757 Info Mar 11 09:04:33 NIM Interface 20 is Link Down

758 Info Mar 11 09:04:38 NIM Interface 20 is Link Up

759 Info Mar 11 09:04:38 NIM Interface 20 is Link Down

760 Info Mar 11 09:04:41 NIM Interface 20 is Link Up

761 Info Mar 11 09:04:54 NIM Interface 20 is Link Down

762 Info Mar 11 09:04:57 NIM Interface 20 is Link Up

As you can see they happen within a few seconds and happen on a varity of equipment (HP PC's, HP notebooks, Printers). I have a support case logged and after 3 firmware upgrades later they are blaming everything but the switcehs. I have had the cabling tested with a Fluke DSX and vertified to Cat5e, upgraded firmware/drivers on PC's, changed patch leads and even changes the switches but i get the same disconnects.

Anyone got a simular problem? Anyone got any idea's?

HP are going down the spanning tree route and blaming the topology changes on another switch (this i am working on) but i don't see how spanning tree can be affecting so many devices on switches that don't support spanning tree?

Please help!

Kris

That looks normal - start of day, people arrive at work and boot up their PC, switchport comes up.

Thanks for your answer.

Sorry that might have been a bad example as they are all in the morning, but these disconnects happen throughout the day on many devices, some of which we know have not been shut down.

Also if it was a simple plug in/turn on why does the port connect/disconnect 4 or 5 times within 30 seconds? Unless the switch does something special when deivces are plugged in.

I have been looking at this with HP for about a month now, replaced switches, updated firmware etc but it still does it. The reason i'm posting to here is the customer does experience some weird network problems and i would like to know if this is a contributing factor.

Kris

a. When PCs go to sleep, interfaces go down. Looks normal.

b. The multiple lines being logged is due to the auto-negotiation taking place, followed by the STP operation. Looks normal.

What weird netwopkr issues.

Thanks again for the reply.

That switch does not have STP as it is a 48 port and not a V2 switch. I get what you are saying about auto neg, i thought that might be that myself, however was a little confused as it seemed to happen during the day on PC's that we have switched power saving off-on.

They are getting random network slow/problems accessing the server. I have a Fluke Optiview network diagnostic which i left there for a week and low and behold the problem didn't occur during that week. It's very sporadic. I have Logic Monitor monitoring the server and does not report any issues. I really am struggling to pinpoint the problem.

They have lots of wireless devices that connect to various monitoring devices, but i have vlan'd them all off to a separate network to eliminate that as the problem.

Really don't know where else to go.

Does everybody complain of the issue simultaneously, when it happens?

Find whoever is complaining and put two .bat file on their desktop that will ping the server and pipe the results to two .txt files.

Tell them, next time it happens, to doubleclick on both .bat files and then call you.

Make the first .bat file ping their default GW (or some other server, eg the AD server or something), the 2nd .bat file pings the server.

This is good PR as your users get the feel you are doing something positive, and it also engages them in the troubleshooting process.

I have no doubt both pings will show no congestion or dropped pings and you will be able to conclude that the network isn't causing the slowdown, but the server is to blame.

Hi to everbody,

a customer of mind has the same problem with efforts on the network.

Databases and Outlook are crashing on a regular basis up to 5 times a day.

Has anyone a solution? Thanks.

On PCs, on NIC properties, disable all power saving modes.

On switch, when using SpanningTree use admin-edge or portfast modes