Re: Weird Network Connectivity Problem
- From: carl.sawyer@xxxxxxxxx
- Date: Wed, 12 Nov 2008 21:47:12 -0800 (PST)
Hi Dave,
Thanks much for your reply and suggestions. Sorry it's taken me a
while to reply, but here you go:
I would immediately replace the cable between the workstation and the wall,All have been done, long ago and many times since. In fact, since
and also the one between the patch panel and the switch. At the same time,
choose a different switch port than the one you've been using.
this first started happening we have replaced the switch.
make sure your NIC driver on the workstation is current, and run the NICHave changed the NIC twice. Have even switched NICs while the x199 is
diagnostic app if one exists. If the app fails the diagnostic, or if the
problem comes back after having replaced the cables and trying a different
switch port, I'd replace the NIC.
in limbo, booted up and it's still broke. I log in remote after hours
and now it's working fine.
Based on your description of an issue that happens rarely, and that alsoIt can't be hardware. When this happens x199 can ping anything on the
seems to spontaneously fix itself, this is almost certainly hardware
related. You don't have a setting that's changing itself every few months,
then changing itself back.
local network EXCEPT the SBS server, the server can ping anything
EXCEPT x199 and everything else can ping BOTH of them.
please resist the temptation to mess withToo late. I've changed every piece of hardware except the workstation
settings that are working almost all of the time.
and the server. I've changed NICs, workstation IP, all the simple
"settings". What I haven't done is remove the computer account in the
domain or start playing around in AD to see what happens. I didn't
think I'd get a solid diagnosis from my post but I'll take any clues I
can get and apply them prudently.
Lemme summarize - Fred logs in to x199, some time later x199-to-server-
to-x199 goes into limbo. Maybe this happens once in three months,
maybe it happens four or five times in one day and not again for a
month and a half. Whenever it happens it eventually "heals". I've
never seen a pattern until last week when it came back at the same
instant that fred was being added to the local administrators group on
another machine, the INSTANT that "Check Name" was clicked. This
reminded me of another time when the problem disappeared when fred
logged in on another machine. Pattern? Maybe the other times
something else happened with the fred account that was not noticed.
Or maybe it's just an amazing coincidence. BYW, I have no idea
whatsoever as to what may be causing the problem in the forst place.
My uneducated quess is "AD weirdness" but that's not a very specific
diagnosis. It's like blaming it on bad karma or the evil eye (I'll
consider any possibility!).
And here's the bitch - I can't make it break so I can't pick a time to
fix it. It breaks when fred is busy (always) and I'm at the other end
of a remote login. And then it heals while I'm pinging and doing
nslookups and comparing arp info. So I'm just tossing a net out to
see if I can catch any ideas. I've been doing networks since the
Datapoint days of the late 70's and I've been doing PC networks since
LANMAN and Novell 2.something but my deep expertise is Unix (I helped
develop the first TCP/IP course taught by good ol' SCO and so forth)
and this has got me (and every person I've described it to) completely
turned around. The hardware is right. The switch is right. The IPs
and the netmasks and the DNS settings are right. Every other machine
in the place is fine. It's GOT to be something confusing AD. Maybe.
I mean, I can purge the arp table on the server, ping x199 by IP and
it times out yet the arp table is updated correctly. WTF?
Sorry for going on. To tell the truth, I can't WAIT to figure this
one out because, no matter what the cause, I'm gonna learn something
before this is done. One of the best networking dinks I've ever known
told me he was glad this wasn't HIS customer!
OK, I'll shut up. Thanks again for your suggestions but I'll see you
and raise you. Any other ideas?
C
On Nov 7, 11:15 am, "Dave Nickason [SBS MVP]"
<gwdib...@xxxxxxxxxxxxxxxxxxxxxx> wrote:
I would immediately replace the cable between the workstation and the wall,
and also the one between the patch panel and the switch. At the same time,
choose a different switch port than the one you've been using. I would also
make sure your NIC driver on the workstation is current, and run the NIC
diagnostic app if one exists. If the app fails the diagnostic, or if the
problem comes back after having replaced the cables and trying a different
switch port, I'd replace the NIC.
Based on your description of an issue that happens rarely, and that also
seems to spontaneously fix itself, this is almost certainly hardware
related. You don't have a setting that's changing itself every few months,
then changing itself back. There's obviously no guarantee that replacing
the cable and the NIC will fix this, but it seems like doing so would cover
almost all of the bases at a low cost. In the absence of a solid diagnosis
that leads you to the contrary, please resist the temptation to mess with
settings that are working almost all of the time.
<carl.saw...@xxxxxxxxx> wrote in message
news:6bb922b2-b825-4076-bcdd-7ad85f961083@xxxxxxxxxxxxxxxxxxxxxxxxxxxxxx
I have a very weird problem that I have been fighting for many, many
months
and I'm looking for any hint, any idea, any possible lead. Here's
the
situation.
● Small network - Windows SBS 2003 server and 15-18 client PCs, all
XP. We
also have a Terminal Services server and a Unix box. All workstations
have
static IPs.
● One machine, \\x199, has an odd and intermittent problem. Every
few
months it will spontaneously and seemingly randomly drop off the
network. It
also will randomly reconnect. This may happen for three or four days
in a
row and then not happen again for two or three months. It just
happened a
few minutes ago and I'll use this incident to illustrate:
1. User "fred" is working away, connected to Outlook, using shares and
apps
on the server, no problems.
2. Suddenly Outlook says it cannot connect. Shared drives are
unavailable.
Shortcuts to server apps are dead.
3. I log into the SBS server (remote).
ping x199 returns "request timed out".
ping 192.168.1.110 returns "request timed out"
When this has happened before I have tried:
ping -a 192.168.1.110 resolves the hostname but returns "request timed
out"
nslookup x199 resolves correctly
arp -a returns correct information
...and from the workstation a ping to the server times out but the
workstation
can ping any other machine in the place, it can see the gateway, and
it can
browse the Internet except that the SBS server also provides DNS. If
we give
it another DNS server or browse to an IP it works fine.
4. I log into our Unix box and ping 192.168.1.110 and it replies w/no
problems.
5. I log into the Terminal Services server and I can ping the
workstation by
name or IP and I can even log into it using Remote Desktop.
6. While this is going on someone on site is setting up another
machine
(\\x116) for fred to work on. Part of this process is to make fred's
domain
account a member of the local Administrators group. At the instant
that they
clicked on "Check Name" on \\x116 fred's Outlook client on \\x199
connected
to the server and my ping -t from the SBS server began responding.
Really.
Once before the workstation was experiencing this problem and fred
logged on
to the machine at the next desk and at the same moment \\x199 started
responding; we thought it was coincidence.
In this case \\x199 dropped off while in use. In other cases the user
tries
to log in at the beginning of the day and discovers that the machine
is
offline.
One of the frustrating aspects of this problem is that it won't stay
broke!
I might be in the middle of testing and suddenly it just starts
working.
Before I discovered that the connectivity issue was only with the SBS
server
we have been through all the usual stuff to diagnose this problem:
different
cable, different NIC, change ports on the switch, etc. To me the
confounding
issue is that the problem is only between the SBS server and this
single
workstation. It would take hours for me to go back through my old
notes and
develop a comprehensive list of everything we've done but if I recall
correctly the problem is not just related to fred's account. When
diagnosing
this problem we have logged on to \\x199 as administrator while it
was
offline and it made no difference.
I have considered stuff like reloading the machine, removing the
computer
account in the domain and recreating it, or driving a wooden stake
through
the machine and burying it at a crossroads but I would prefer to have
some
idea as to what's wrong before I spend time just shooting in the
dark. Any
ideas? Suggestions? Incantations or exorcism ceremonies I can try???- Hide quoted text -
- Show quoted text -
.
- References:
- Weird Network Connectivity Problem
- From: carl . sawyer
- Re: Weird Network Connectivity Problem
- From: Dave Nickason [SBS MVP]
- Weird Network Connectivity Problem
- Prev by Date: RWW Certificate issue
- Next by Date: Re: Unable to view the performance report
- Previous by thread: Re: Weird Network Connectivity Problem
- Next by thread: GPO fast user switch?
- Index(es):
Relevant Pages
|