Re: (ASCEND) COMPLETE BULLSHIT

[matt sommer <list9@radha.webkorner.com>]

Troy Settle wrote:
> 
> ** it's a software bug--happens to 4048's too.
> 
> I *seriously* doubt it.  Out of 11 boxes running 7.0.22, this is the only
> box it's happened on.

then it is hardware?  did ascend rma it?

 
> What's different about this box than all the others?
> 

for me the difference was the load.


> For starters, it has a 10/100 ethernet port.  Ok, that's 3 of the 11.
> 
> Now, of those 3, 2 are connected to a 10/100 switch running in 100MB / Full
> duplex.  These 2 6000's haven't had their ethernet go down, but the Rx and
> Tx stats are way off.  (100 million packets recieved, only 50k packets
> transmitted???).

and that would be broken hardware or software???

are you sure the config is identical--arp, ospf, pools, masks, etc? 
have you tried swapping boxes?


> The remaining max is on a 10mbit hub, and is having problems.
> 
> There is only 1 warning in the fatal history, and it's not in reference to
> anything related to ethernet (so Ascend says).
> 
> All this leads me to believe that it's a severe problem with the packet
> and/or hardware drivers for the ethernet interface on the 6000 chassis.

ok, but that would be software--something that is fixed by an upgrade
(or downgrade).

so maybe the driver configures slightly differently for 10b-T on 10/100
interface--maybe a factor...  maybe you discovered a new way to break
the software--another bug requiring another patch--a patch which may not
do anything to strengthen the basic architecture.

(forgive all my random speculation--i know i don't know.)

if it is software, they will never tell us what it is exactly, and we
shouldn't have to care.  as you say, it *is* a severe problem.  and
complete bullshit?  arguably so, considering their release history...


> I gave Ascend your name, and they looked up the ticket in question, and said
> it was an unrelated problem (you were actually seeing errors logged in your
> fatal history).

unrelated, ya --  ;)    ("that was issue #ewww4xyz -- this is issue
#ewww4pdq")  slightly different symptoms, yes, but a dead interface is a
dead interface.  and i mean *dead*--not even arp coming out.  same for
you?  for me, they have already *proven* they can break the interface
with bad software.

if it is broken hardware and they told you so you should tell us--but
you say "go into Ethernet -> Mod Config -> Ether Options, and change
something... ANYTHING, and *BOOM* the etnernet starts working again." 
that tells me your hardware is not broken.  maybe a buggy ASIC that gets
reinitialized when you save a change?  possible, but the likely suspect
is their software, right?  as you said, probably a weak, easily broken
layer two driver.  

anyway, not our department, they just need to fix it--fast.  what did
they advise you to do anyway?  change config?  go to 7.2.x or 7.4.x
software?  rma?  (what release are most of you trusting on 6096 anyway?)

 
> The engineer I spoke with said that the fix in 7.0.21e4 was indeed carried
> over to 7.0.22.  

i hope he is right, but of course sometimes they feed you "(ASCEND)
COMPLETE BULLSHIT".

wanna try a copy of 7.0.21e4?
++ Ascend Users Mailing List ++
To unsubscribe:	send unsubscribe to ascend-users-request@bungi.com
To get FAQ'd:	<http://www.nealis.net/ascend/faq>