[Discuss] connection issue

[jdm at moylan.us (dan moylan)]

doug mildram writes:
> dmildram>
> I'm admiring all the Q's and ideas on your networking
> problem.  John Abreau's workaround idea using eg VONET $26
> wireless bridge to put wired device(s) into wireless use was
> enlightening to me for general home use..., even though I
> got a bit confused on the overall setup he used and how that
> might help wireless clients see each other.
> Tangent/useless: If WIRED clients were ganged together on a
> mini switch, that would usually solve  "clients not able to
> see each other" !  Obviously.

> dmildram>
> ANYWAYS, ie back to Dan Moylan's quest for WhatsGoingOn :
> Symptoms sound much like what a PVLAN does.   From
> wikipedia: *Private VLAN*, also known as *port isolation*,
> is a technique in computer networking where a VLAN contains
> switch ports that are restricted such that they can only
> communicate with a given uplink.
>
> dmildram>
> I ran into this layer2 feature in (odd product, to me) EMC
> Centera: one cluster per rack w/many 4-disk 1U boxes AND 2
> enetswitches.  Weird storage.  I worked at EMC many months
> maintaining a lab of Centeras, before realizing that within
> each rack, most nodes were purposefully isolated from each
> other, thus introducing me to the rarely-used
> Private/restricted-port PVLAN concept.  After that strange
> world, I got a much-better-learning networking job.  Pardon
> TMI.

> dmildram>
> back to Dan and home use/problem: Lord knows these
> all-in-one HOME "router" boxes ( AP + switch + router ) keep
> users less aware of internals, so I wonder how/if PVLAN
> could have come to life in a t-mobile router+switch....a
> possible security feature...  since you said it worked fine
> up until a week or so ago!  So while I doubt it, I still
> Hope This Helps mentally or better.  Sure is hard to "try
> plugging all into this dumb switch together" for wifi !
> p.s. also: I would be inspecting ARP tables though with
> *nix#  arp -a ....if ports are isolated, only the client
> with the target IPaddr would reply to an ARP request    (
> request/re: the target ipaddr ) if I'm not mistaken.
> Finally,  doublecheck that 192.168.0.x has a  /24
> (255.255.255.0) NETMASK (  various show-me cmds like
> "ifconfig -a" or "ip a",  or windows "ipconfig /a" ) but
> offhand I'd wager anything using 192.168.0.x  keeps the std
> /24 mask.

> original/early problem post w/o many ideas/replies:

>>* dan moylan wrote:

>>>* the problems keep mounting and puzzling (moan).  computers
>>>* alphacent, aldeberan (both fc36) and rigel (fc27) all show
>>>* as connected on the local t-mobile wireless app, and locally
>>>* with ifconfig.  route shows 192.168.12.0 gateway on all
>>>* three.  they can all ping blu.org <http://blu.org>
>>>* successfully, but not each
>>>* other, nor can they ssh into each other.  iirc everything was
>>>* fine a few days ago.  what have i done?  what's going on?*

i went off to talk to my friend at t-mobile about this,
taking alphacent with me, and he recommended that i just
reboot the t-mobile wifi access device.  (should have
thought of that myself).

when i got home, before rebooting the t-mobile wifi device,
alphacent wouldn't come out of hibernation, so i rebooted
it.  when i did, the wifi port isolation issue had
disappeared.  sadly, previously i had tried pinging from
alphacent to the other two, and from each of the other two
to alphacent, but i don't believe i had tried pinging from
aldeberan to rigel (dumb oversight).  now, i can only wonder
(though i think i know).  #@*&^!%$@  at least the problem
has now gone away.  can anyone tell me what the issue likely
was in aphacent or how i could have detected that?

tia,
ole dan

_______________________________________________
Discuss mailing list
Discuss at lists.blu.org
http://lists.blu.org/mailman/listinfo/discuss