ra 3.0.0 client vs 2.0.6 Argus data - Bug?

[Carter Bullard]

Hey Jesper,
I suspect that you're ra() is exiting correctly somewhere early in the  
data file.

If you don't mind doing some debugging, if you configure and compile
your clients with the tag files ".debug" and  ".devel" in the root  
directory,
and then running ra() with some debug turned on, that should tell us
if something is up, and then use gdb to see the actual error.

    % cd /path/to/argus/clients/directory
    % touch .devel .debug
    % make clobber
    % ./configure;make

    % bin/ra -D3 -r /path/to/argus_2.0.6_new

The last debug strings should reveal if there is a real reason to exit.
If there is nothing telling, increasing the "-D num" number will tell
us more.

ra* clients can exit if they receive specific management records, such  
as
those that indicate the stream has errors, or has closed.  I suspect  
that
a 2.0.6 MAR record is telling the 3.0.0 ra() to close, but it doesn't  
have
to.

Carter

On Jan 30, 2009, at 8:13 AM, Jesper Skou Jensen wrote:

> Yes, on two different Argus boxes. 2.0.6 on the old box and 3.0.0 on  
> the new.
>
> The ra mentioned below is the the 3.0.0 version and it's reading the  
> 2.0.6 logs.
>
>
> -- 
>
>  Jesper S. Jensen
> UNI-C - Århus, Danmark
>
>
> Dave Edelman wrote:
>> Do you have a mix of 2.0 and 3.0 clients?
>> Dave Edelman
>> +1 917 331-0112 cell
>> On Jan 30, 2009, at 4:12, Jesper Skou Jensen <jesper.skou.jensen at uni-c.dk 
>> > wrote:
>>> carter at qosient.com wrote:
>>>> Hey Jesper,
>>>
>>> Hi Carter,
>>>
>>> Thank you for the quick answer.
>>>
>>>> The data model in 3.0.0 is much different, and so the size of the  
>>>> file, etc, is not a good metric for comparison.
>>>> I use racount() to compare if the basic contents are the same.
>>>
>>> Ok, thanks for the info.
>>>
>>>> The situation where you read a 2.0.6 file and generate 2.0.6_new  
>>>> looks to be the one that has big problems, as it looks like you  
>>>> are missing 90% ?
>>>
>>> Yea, that's quite a big problem imo, and yea there is missing  
>>> quite a lot of data.
>>>
>>> > Is ra() vaulting?
>>>
>>> Do you mean if it's crashing or writing any garbage or something  
>>> like that? No. Not that I know of. The output I posted was/is all  
>>> it gives.
>>>
>>> racount confirms what wc did, the _new logfile only contains a  
>>> fraction of the original log.
>>>
>>> If I do eg:
>>> ra -r argus_2.0.6.argus |tail
>>>
>>> I do get the last entries in the log, so the read part appears to  
>>> be running ok. It must be the write that that goes wrong somehow?
>>>
>>>
>>> -- 
>>>
>>> Jesper S. Jensen
>>> UNI-C - Århus, Danmark
>>>
>

Carter Bullard
CEO/President
QoSient, LLC
150 E 57th Street Suite 12D
New York, New York  10022

+1 212 588-9133 Phone
+1 212 588-9134 Fax