BUS ERROR !!!!!

[gfunk123]

Hi there

I am running soalris 9 on a sun fire 480r and all of a sudden (today) whenever the users run the command `top` we get the following message `bus error`

does anybody have any information on what this is all about and whether there is a routine i can perform to gather more information on this

any help on this would greatly appreciated

 

[dblair]

start by looking at your system warning/error logs.
likely to be a SCSI error/timeout.

 

[gfunk123]

im running a tail -f /var/adm/messages and in another window i am generating the error. However , nothing is being logged in the messages file

baffled ! What does bus error actually mean? is it an overload of an internal data bus (ie physical bus) or is it a software based error ?

 

[dblair]

Can you post the xcat error.
do you only get this when running top ??

 

[gfunk123]

sorry can you explain how i get the `xcat error` im not familiar with this

cheers

 

[vgersh99]

'xcat error' == 'EXACT error'

vlad
+----------------------------+
| #include<disclaimer.h> |
+----------------------------+

 

[Salem]

> What does bus error actually mean?
It's purely a software fault within the application which generates the bus error. This is usually treated as a fatal error by the operating system, and the program is halted.

Since it seems to be a new problem, what else has happened to your system. Say patches to your operating system (OS) or other software installed. Since top needs to examine the OS at a fairly detailed level, perhaps top is a patch level behind the rest of the OS (talk to your support provider if this is the case).

As for what you can do about it, the answer is not much. If there is a real problem, then it needs to be fixed by the author of the program fixing the source code and recompiling the program.

There is one piece of information which may be of interest to a developer, and that is a core file. This is an exact snapshot of the memory of a program at the point it failed, and can be examined by a software developer to determine the cause of the problem.

If your user profile is set up to prevent core files being produced, you should be able to enable it using the following csh commands.

$ limit coredumpsize unlimited
$ top

When the top command crashes with a bus error, you should find a file called 'core' in your current directory.
Mention that you have a core file in your support query. Core files are large, so you wouldn't want to send it until it was specifically requested.

--

 

[gfunk123]

exact error is

#> top
bus error



thats it !!

 

[gfunk123]

ah, ive now discovered it is completely intermittant, I just tried top and it worked, then i tried again and it said `bus error`


this is crazy

 

[gfunk123]

no core files at / but we use ksh profiles and there is no line to specifically restrict core dumping (unless it is the default?)

are you saying that it would potentially coredump every time the error is generated? cos that would be a lot of core's cos i have a whole dept trying to use it at the mo

In answer to your question salem, there have been absolutely no chanes made to this box in the last 6 months (i am only administrator) so no patches , software installations etc

 

[dblair]

why is everyone using top.. This may actualy cause the bus error message..

 

[gfunk123]

long story, theres actually about 4 people use top (need to keep constant check on their processes....you say this could be causing the error,can you explain how ???