Google Groups no longer supports new Usenet posts or subscriptions. Historical content remains viewable.
Dismiss

huge collision and bad frame alignment

0 views
Skip to first unread message

johnny lo at jvb

unread,
May 20, 1999, 3:00:00 AM5/20/99
to
This is the config of our server.

One NW411 server
One 3C905B segment
One Seagate STD28000N internal tape drive
One CD ROM IDE
About 25 stations with a mix of mostly 10Mb and a few 100Mb on a 10/100Mb
autoswitch hub.

Problem: The console monitor screen, there are huge number of collision,
send deferred and send abort especially in the morning. There are some "..
bad frame alignment" error and yesterday it bump right up to 6000+ from
below 10. There was a 20min windows yesterday morning that some stations
can't access the server. During the time only less than half the staff was
in and they were DOS stations. I am suspecting a very long BNC extension
connecting a hub in a department in the other corner of the building. But
wouldn't it be problematic all the time instead of sometimes?

Any suggestion is appreciated.

J

johnny lo at jvb

unread,
May 20, 1999, 3:00:00 AM5/20/99
to

Brad Doster

unread,
May 21, 1999, 3:00:00 AM5/21/99
to
In article <7i2c23$88...@cronus.novell.com>, Johnny lo at jvb wrote:
> I am suspecting a very long BNC extension
> connecting a hub in a department in the other corner of the building. But
> wouldn't it be problematic all the time instead of sometimes?
>
Not necessarily. If there's a loose connector or an internal break in the
cable, moving the cable and/or connector could cause sporadic behavior.

bd
NSC SysOp


Brad Doster

unread,
May 21, 1999, 3:00:00 AM5/21/99
to
Duplicate

bd
NSC SysOp


Frank Carey

unread,
May 23, 1999, 3:00:00 AM5/23/99
to johnny lo at jvb
I have seen similar problems with a cooked NIC (one too many power
blackouts) and a short in 10Base-T wall plate. In the former case, the
card was fine until any workstation tried to login. When a workstation
accessed the server, the error counters zoomed until the server
ABENDed. Swapped the card and fixed the problem. In the latter case we
were able to track down the problem by finding a hub port that has a
solid-lit status light. Problem was that there was nothing attached to
that particular port. Disconnect the patch cable for that port and the
problem ended.

Frank Carey

johnny lo at jvb wrote:
>
> This is the config of our server.
>
> One NW411 server
> One 3C905B segment
> One Seagate STD28000N internal tape drive
> One CD ROM IDE
> About 25 stations with a mix of mostly 10Mb and a few 100Mb on a 10/100Mb
> autoswitch hub.
>
> Problem: The console monitor screen, there are huge number of collision,
> send deferred and send abort especially in the morning. There are some "..
> bad frame alignment" error and yesterday it bump right up to 6000+ from
> below 10. There was a 20min windows yesterday morning that some stations
> can't access the server. During the time only less than half the staff was

> in and they were DOS stations. I am suspecting a very long BNC extension


> connecting a hub in a department in the other corner of the building. But
> wouldn't it be problematic all the time instead of sometimes?
>

Jack

unread,
May 24, 1999, 3:00:00 AM5/24/99
to
Thank you for the responses. Now I have narrowed it down a little bit.

Disconnected the long BNC and 10bT cable and leave only 4 hubs on the same
single segment. With just 1 station login (except for print servers, etc),
whenever a search for files is executed on the server volume, collision rate
climbs and stops when the command is finished.

I would start to look at the back bone cables, the cable to the station,
NIC's setting and may eventually replace the NIC. Is SQE setting on the
server NIC worth looking into, and how?

Thanks.

0 new messages