[f-nsp] [SPAM tagged by PCNET] Failing Backplane or Interface Module on a FastIron 800...help is greatly appreciated.

[Jason J. W. Williams]

Hi Guys,

This issue continues to occur. It now occurs on both of our F48-E
interface modules, and seems to be dependent on load. We've checked
the speed/duplex of each port on both the switch and server sides for
all ports on both blades, and found no mismatches. Also, no port on
either switch or server side seems to log any CRC or other frame
errors.

The Foundry SE buddy I talked to said it sounded like CAM corruption,
but I'm having a heck of a time figuring out what's causing it. Any
help is greatly appreciated.

Best Regards,
Jason

On 3/12/07, Jason J. W. Williams <jasonjwwilliams at gmail.com> wrote:
> Hi Todd,
>
> Thank you again. I finally got a straight story out of a Foundry SE.
> You're absolutely correct that speed/duplex mismatches can cause CAM
> trashing. Also, CAM trashing can occur on J-F48E modules in a select
> manufacturing window. Apparently it was a problem with their memory
> supplier. They fixed the issue with the J-F48E-A. Again thank you.
>
> Best Regards,
> Jason
>
> On 3/9/07, Todd Christell <tchristell at springnet.net> wrote:
> >  Jason,
> >
> > The old 24 por 10/100 blades won't work with the JetCore module  .
> >
> > We saw some similar behavior on a BI8000 with JetCore that ended up
> > being a speed/duplex mismatch.  All of our internal ports are set for
> > Auto but one of our media converters went to hard coded 100 meg full
> > dux.  Didn't seem to bother anything until the traffic on that port
> > ramped up to 50 or 60 meg and then the 10/100 blade started wigging out.
> > Dropped traffic, arp tables corrupted and finally the port just died.
> > Not knowing about the media converter we swapped out the card which
> > appeared to fix the problem until the traffic on that port ramped up
> > again and killed the port on the new card.  We replaced the media
> > converter and tested both cards and no problems.
> >
> > Foundry said that the configuration of hard code device to a Foundry
> > port configured for auto-negotiation was an "unsupported configuration."
> >
> > Todd
> >
> > Todd Christell
> > SpringNet Network Manager
> >
> > -----Original Message-----
> > From: foundry-nsp-bounces at puck.nether.net
> > [mailto:foundry-nsp-bounces at puck.nether.net] On Behalf Of Jason J. W.
> > Williams
> > Sent: Friday, March 09, 2007 2:46 PM
> > To: foundry-nsp at puck.nether.net
> > Subject: [SPAM tagged by PCNET] [f-nsp] Failing Backplane or Interface
> > Module on a FastIron 800...help is greatly appreciated.
> > Importance: Low
> >
> > Hello all,
> >
> >
> > I've got kinda of a weird situation, and if you could give me your take
> > it'd really give me some peace of mind. Suddenly last night all of the
> > hosts attached to the 1-24ish ports of a FI48E (JetCore) blade in our
> > FI800 started experiencing massive packet loss (20-100%).
> > Before we figured out it was that blade, we tried failing the active
> > management module (FxGMR4) over to the passive just in case it was the
> > problem. That didn't help. Actually moving the the affected hosts from
> > one FI48E blade to another seems to have corrected the issue. When we
> > ping hosts on the "bad" blade we get stuff like this:
> >
> > wrong data byte #28 should be 0x1c but was 0x7e
> > #8      8 9 a b c d e f 10 11 12 13 14 15 16 17 18 19 1a 1b 7e 1f 1e
> > 1f 20 21 22 23 24 25 26 27
> > #40     28 29 2a 2b 2c 2d 2e 2f 30 31 32 33 34 35 36 37
> >
> > A show backplane and a show module, doesn't show any errors or issues on
> > the "bad" blade.
> >
> > My real concern right now is that we might have a chassis backplane
> > going bad (which would take out everything). Is that possible/likely?
> >
> > Any help in eliminating the backplane as a possibility would really be
> > appreciated. Also, I have some IronCore 24-port blades from our old
> > FastIron II. Could I put those into my two empty slots on the FI800 to
> > take over for the "bad" blade until I can get it replaced? If so, does
> > it require a chassis reboot to do that (should the chassis be off)?
> >
> > Lastly, as an update, some of the hosts connected to the "good" FI48E
> > are starting to exhibit long pings (no packet loss yet). I'm very
> > concerned something's going to take the whole chassis down. We're
> > running on 3 power supplies, instead of 4. Could power be an issue?
> >
> > Again any help is greatly appreciated!
> >
> > Best Regards,
> > Jason
> > _______________________________________________
> > foundry-nsp mailing list
> > foundry-nsp at puck.nether.net
> > http://puck.nether.net/mailman/listinfo/foundry-nsp
> >
>