ILO Problem with Cisco Switch

Solution #2

posted on Aug 11, 2005

Duke - usenet poster

Rank: Apprentice
Rating: 0%, 0 votes

In article <Xns92FF681B13FCAeastburn123 @216.166.71.230>, Marc
<eastburn@<AntiSp@m>speakeasy.­org> says...

Then I would think this is a problem with the mgmt boards.  Or the
drivers for them.  Thanks for the TIP on speakeasy.

--

hsb

"Somehow I imagined this experience would be more rewarding"  Calvin
***************  USE ROT13 TO SEE MY EMAIL ADDRESS  ****************
******************************­******************************­********
Due to the volume of email that I receive, I may not not be able to
reply to emails sent to my account.  Please post a followup instead.
******************************­******************************­********

Was this solution helpful? Show your Appreciation by rating it:

Solution #3

posted on Aug 11, 2005

Brad - usenet poster

Rank: Apprentice
Rating: 0%, 0 votes

Well,

I finally got one of our network folks to look at this, he verified that the
ports were set to portfast and 100/full.

He did mention that he saw some CRC errors and some input errors, but who
knows from when. He cleared the counters and is going to monitor through
tomorrow.

Unfortunately, I do not have access to the switch or the router, I need to
rely on our network team, which up until today has not even taken a look at
the switches in question.

what will the commands that you have listed below show me?

I have not tried to constant ping these mgmt boards but the reason I know
when they go down is that the availability of the device is checked with
SNMP every 10 minutes.

--
==============================­==========================
~~Get Speakeasy Broadband - Where the Customer Rules!~~
      http://www.speakeasy.net/refer /36701

Was this solution helpful? Show your Appreciation by rating it:

Solution #4

posted on Aug 11, 2005

jessie25 - usenet poster

Rank: Apprentice
Rating: 0%, 0 votes

Does Cisco support any features that protect the CPU from being
flooded by shutting down a port that receives too much traffic
destined to the CPU?  Or does it support shutting down a port that
receives too many broadcasts (above a threshold)?

Some vendors have features like that that will cause a port to
shutdown based on certain conditions.  It is not always spanning tree
(comment directed at Jonathan).

--Bernie

Was this solution helpful? Show your Appreciation by rating it:

Solution #5

posted on Aug 11, 2005

Mini Me - usenet poster

Rank: Apprentice
Rating: 0%, 0 votes

Not necessarily...

The switch might be conned into believing that just that port constitutes a
spanning-tree loop... and shutting it down... and when you reset the
connection on the server, the port goes back into forwarding, until again
receiving the bad bpdu...

See above...

Jonathan
BTW, I may be completely wrong... but spanning-tree is generally the only
thing that will shut down a port...

Check the port's status in the IOS the next time it happens...

Was this solution helpful? Show your Appreciation by rating it:

Solution #6

posted on Aug 11, 2005

Luisa_K - usenet poster

Rank: Apprentice
Rating: 0%, 0 votes

Sure, go here:

http://cgi.cs.wisc.edu/scripts /ballard/bofhserver.pl

Hit reload a bunch of times after you get there...

Jonathan

Was this solution helpful? Show your Appreciation by rating it:

Solution #7

posted on Aug 11, 2005

Lizzy - usenet poster

Rank: Apprentice
Rating: 0%, 0 votes

In article <Xns9302741958F9Beastburn123 @216.166.71.230>, Marc
<eastburn@<AntiSp@m>speakeasy.­org> says...

<G>  Maybe they verified it *after* they changed it for you!  

--

hsb

"Somehow I imagined this experience would be more rewarding"  Calvin
***************  USE ROT13 TO SEE MY EMAIL ADDRESS  ****************
******************************­******************************­********
Due to the volume of email that I receive, I may not not be able to
reply to emails sent to my account.  Please post a followup instead.
******************************­******************************­********

Was this solution helpful? Show your Appreciation by rating it:

Solution #8

posted on Aug 11, 2005

Bouncy - usenet poster

Rank: Apprentice
Rating: 0%, 0 votes

Hansang Bae <u @alp.ee.pbz> wrote in

Well that was my original thought, but now I'm not sure. My network guy
finally "Verified" that the ports were set to 100/full and portfast, and
curiously, the Mgmt borads did not go down over the weekend. I'm going to
give it another few days, but since this started happening over a month
ago, they have gone down every 2 days without exception.

My only guess is that the settings were changed on the switch and I was not
told.

Was this solution helpful? Show your Appreciation by rating it:

Solution #9

posted on Aug 11, 2005

Duke - usenet poster

Rank: Apprentice
Rating: 0%, 0 votes

We had an SLA with Ameritech (The Big Evil in Illinois...), and every time
you would call them about a line being down, they would put you on hold and
then say "I see nothing wrong..."

And, of course, in that time the circuit came up...

Yeah... uh huh... sure.

Jonathan

Was this solution helpful? Show your Appreciation by rating it:

Solution #10

posted on Aug 11, 2005

Charlie - usenet poster

Rank: Apprentice
Rating: 0%, 0 votes

In article <hlCdnXDMs_gLpYOjXTW @giganews.com>, "Eastburn" <eastburn@
<antisp@m>speakeasy.org> says...

It'll tell you when you last cleared the counters at the bottom.

Always a good idea.

Those Network Bastards! :)  

They will tell you if there are any suspect error messages (sho log
buffer 1023), "sho cam dyn mod/port" will tell you what MAC address was
learned or more importantly if it disappears, and - on the router - "sho
ip arp MGMT_IP_HERE" will tell you if the router is losing the arp
table.  Keep an eye on the MAC address as someone may be hijacking the
IP address.

I guess that's often enough.  I just wanted to make sure no one was
falling asleep at the wheels - sorta speak.

How do you like Speakeasy by the way?

--

hsb

"Somehow I imagined this experience would be more rewarding"  Calvin
***************  USE ROT13 TO SEE MY EMAIL ADDRESS  ****************
******************************­******************************­********
Due to the volume of email that I receive, I may not not be able to
reply to emails sent to my account.  Please post a followup instead.
******************************­******************************­********

Was this solution helpful? Show your Appreciation by rating it:

Solution #11

posted on Aug 11, 2005

Odud - usenet poster

Rank: Apprentice
Rating: 0%, 0 votes

Hansang Bae <u @alp.ee.pbz> wrote in



Back to my original post  .. this is happening to 10 different devices
plugged into 3 different switches. All 10 are of the same model, we have
other management boards that are different models, and we have no problems
with them at all (they are plugged into the same switches.)

I really hope that I don't have 10 IP's that are getting Hijacked every 2
days.

It's great, a little bit of downtime when I first got it 2 years ago, but
since then very few problems, Static IP's, Installation was painless, I can  
(and do) run a private network including servers with the blessing of
speakeasy. I also just upgraded to 1.5/128 service (tested at
dslreports.com at about 1.25/90), it is awesome, I might just bump the
upload speed to 384 in the future.

I also like that you can add to your service from a menu, need a second IP,
great, just buy one, how about dial up, sure buy that too.

Was this solution helpful? Show your Appreciation by rating it:

Solution #12

posted on Aug 11, 2005

herself - usenet poster

Rank: Apprentice
Rating: 0%, 0 votes

I am far from an expert but it sounds to me like there is a conflict or
incompatibiliby between the loa and the cisco.  I dont think your test
gets HP (compaq) off the hook.

try pluggin all the ilos in to dumb hubs and the hubs into the switches.
or the ilos into some other brand of switch.

and get back on the phone to HP (compaq) and escalate the problem till you
get to someone that can help you.  dont take NO for an answer.

bill

Was this solution helpful? Show your Appreciation by rating it:

Solution #13

posted on Aug 11, 2005

Janice - usenet poster

Rank: Apprentice
Rating: 0%, 0 votes

"Jonathan" <jrcd @nospam.nospam.hotmail.com> wrote in



It's not actually the NIC in the server it's the Embedded NIC in the
Remote Insight Board (sort of a remote console device) that keeps
dropping, server NIC's are fine (connected to different switch). I'm
running a test right now where I have plugged the Insight board into the
same switch as the servers  .. we will see what happens. Also there are
other servers connected to the same switch as the Insight boards.

Checked out power management ... that's not it.

I also forwaded some information that I found on Cisco's website to the
Network guys here ... Hope they read it.

http://www.cisco.com/warp/publ ic/473/46.html

Was this solution helpful? Show your Appreciation by rating it:

Solution #14

posted on Aug 11, 2005

Mini Me - usenet poster

Rank: Apprentice
Rating: 0%, 0 votes

Wouldn't it be cool if it did?

In a broadcast storm, the switches peak utilization and eventually crash...

However, Cisco may have fixed this... hb???

I was just speculating on what would be causing a switch port to disable...
the number one cause would be detecting a spanning-tree loop...

However, I hadn't thought about errdisable...  (which in retrosepect seems
the more likely cause... the odds that the servers were spitting out bpdu's
is remarkably low... But I could conceive MS thowing spanning-tree on their
machines... )

Jonathan

Was this solution helpful? Show your Appreciation by rating it:

Solution #15

posted on Aug 11, 2005

lawyer - usenet poster

Rank: Apprentice
Rating: 0%, 0 votes

In article
<C69EA17CE0F92982.EB9C97B5359F­C0B4.E16DCD3BF2C1C @lp.airnews.net>,
Ber @weekend.com says...

Nothing for auto-detecting traffic bound for the switch.  But the
broadcast
level can be handled via "set port broadcast mod/port_range 50%" or
whatever percentage you want.

--

hsb

"Somehow I imagined this experience would be more rewarding"  Calvin
***************  USE ROT13 TO SEE MY EMAIL ADDRESS  ****************
******************************­******************************­********
Due to the volume of email that I receive, I may not not be able to
reply to emails sent to my account.  Please post a followup instead.
******************************­******************************­********

Was this solution helpful? Show your Appreciation by rating it:

Solution #16

posted on Aug 11, 2005

Pasty - usenet poster

Rank: Apprentice
Rating: 0%, 0 votes

"Jonathan" <jrcd @nospam.nospam.hotmail.com> wrote in



From what I have read about errDisable, it should change the led on the
port to orange and would also need to be manually reset on the switch, so
simply resetting the end device would not get it back on the network. The
led's for these ports do not go orange, they are green, and the link
light on the device is on.

Is my understanding of errDisable correct?

Was this solution helpful? Show your Appreciation by rating it:

Solution #17

posted on Aug 11, 2005

2Pansy - usenet poster

Rank: Apprentice
Rating: 0%, 0 votes

In article <Xns92FE67592DECeastburn123 @216.166.71.230>, Marc
<eastburn@<AntiSp@m>speakeasy.­org> says...

Yes.  So there's nothing in the switch log?  "sho log buffer 1023" ?  
It just stops responding?  What does "sho ip arp IP_HERE" say on the
router?  What does "sho cam dyn mod/port" say?  (mod/port of the mgmt
board)  What happens if you ping the mgmt every 5 minutes?

--

hsb

"Somehow I imagined this experience would be more rewarding"  Calvin
***************  USE ROT13 TO SEE MY EMAIL ADDRESS  ****************
******************************­******************************­********
Due to the volume of email that I receive, I may not not be able to
reply to emails sent to my account.  Please post a followup instead.
******************************­******************************­********

Was this solution helpful? Show your Appreciation by rating it:

Solution #18

posted on Aug 11, 2005

pandamama - usenet poster

Rank: Apprentice
Rating: 0%, 0 votes

I would have thought that Compaq would have picked that one up pretty
quick...

It also seems improbable that one would set up power management on the
NICs... most PC/Servers, keep power to the NIC even when the machine is off
(to facilitate WOL)... also to ensure that the switchport is still active so
spanning-tree doesn't prevent logins/bootp/etc... so that you do not need to
enable portfast to maintain network reachability... (this way, if you have
multiple NICs in a Server for fail over (hot standby IP, where both NICs are
up and powered and share a MAC address... I have seen this in clustering
environments) and some idiot enables IP forwarding, you don't accidentally
shut down your network...)

Jonathan

Was this solution helpful? Show your Appreciation by rating it:

Solution #19

posted on Aug 11, 2005

kcw573 - usenet poster

Rank: Apprentice
Rating: 0%, 0 votes

I'd like to be the one that checks the port settings, unfortunatly, I
dont have access to the switch, I'm having trouble getting our network
engineering team engaged with this. They beleive that this could not be a
problem with the switch, one engineer actually guaranteed that the switch
was not the problem ... I asked for that in writing and he backed off,
but I'm still not getting much help. The only thing that they have done
for me is verify port speed and IOS ...

Right now I have 10 different servers on 3 different switches with this
problem.

Can you point me to an article or config guide from Cisco that explains
this problem so I can forward it to  him and his boss.

Thanks

Marc

"Jonathan" <jrcd @nospam.nospam.hotmail.com> wrote in

Was this solution helpful? Show your Appreciation by rating it:

Solution #20

posted on Aug 11, 2005

Gary10 - usenet poster

Rank: Apprentice
Rating: 0%, 0 votes

This is a guess...

There is only one reason a switch port will move from the forwarding to the
blocking state... and that is if it discovers a spanning-tree loop...

But, since we do not know what the switch is actually doing, we can only
guess here... it may actually be a problem with the NICs in your servers...
and Compaq is playing the "it must be someone else's problem, because we
don't know how to fix it..."

Jonathan

Was this solution helpful? Show your Appreciation by rating it:

Solution #21

posted on Aug 11, 2005

Ranny - usenet poster

Rank: Apprentice
Rating: 0%, 0 votes

You really need the assistance of your network guy. Spanning tree would not
normally shut down a port permanently. The switch should have "spanning tree
portfast" enabled so that it starts forwarding as soon as the link comes up
and to eliminate speed/duplex negotiation have it's port fixed at 100M FD.

Was this solution helpful? Show your Appreciation by rating it:

Solution #22

posted on Aug 11, 2005

Cornish - usenet poster

Rank: Apprentice
Rating: 0%, 0 votes

In article <epZS9.152545$iw4.20 @news.easynews.com>,
jrcd @nospam.nospam.hotmail.com says...

Or it could be shut down due to errdisable.  I.e. too many errors on the
port.

Like power management accidently putting the NICs to sleep, for example.

--

hsb

"Somehow I imagined this experience would be more rewarding"  Calvin
***************  USE ROT13 TO SEE MY EMAIL ADDRESS  ****************
******************************­******************************­********
Due to the volume of email that I receive, I may not not be able to
reply to emails sent to my account.  Please post a followup instead.
******************************­******************************­********

Was this solution helpful? Show your Appreciation by rating it:

Solution #23

posted on Aug 11, 2005

Rachel007 - usenet poster

Rank: Apprentice
Rating: 0%, 0 votes

It sounds like spanning tree is blocking those ports.... So, either the
Compaq's are generating Bpdu's (possible...) indicating a ST loop, or the
network adapters on the Compaqs are shutting down...

Maybe:

On the Cisco Switch ports:

spanning-tree portfast

Was this solution helpful? Show your Appreciation by rating it:

Solution #24

posted on Aug 11, 2005

Mini Me - usenet poster

Rank: Apprentice
Rating: 0%, 0 votes

Thanks Jonathan,

One quick question, if Spanning Tree is the culprit, wouldn't it also case
this problem on other devices too.

I have several other devices plugged into the switches and do not have
problems with those.

Thanks again

Marc

--
==============================­==========================
~~Get Speakeasy Broadband - Where the Customer Rules!~~
      http://www.speakeasy.net/refer /36701

Was this solution helpful? Show your Appreciation by rating it: