possible radium issue
Phillip G Deneault
deneault at WPI.EDU
Sun Jul 12 21:02:07 EDT 2009
Sorry for the late reply, I tested it today and was successful. Only one
srcid per file now.
I will agree, I didn't notice a problem with rastream. This only seemed
to occur in rasplit.
Thanks!
Phil
On Thu, 9 Jul 2009, Carter Bullard wrote:
> Hey Phillip,
> I refreshed the argus-clients-3.0.2.beta.9 with fixes for the rasplit()
> problem.
> rastream() is ok on this one right now. Grab the image and give it a quick
> try.
>
> http://qosient.com/argus/dev/argus-clients-3.0.2.beta.9.tar.gz
>
> Carter
>
> On Jul 9, 2009, at 1:57 PM, Phillip Deneault wrote:
>
>> I can confirm every file being generated has multiple srcid's in them
>> despite having the $srcid criteria in the rasplit lines in all three time
>> slices.
>>
>> Thanks,
>> Phil
>>
>> Carter Bullard wrote:
>>> The tell tale symptom would be that you have slices with multiple
>>> srcid's in them, even though your output filename has the
>>> $srcid directive in it.
>>> I'm working this now, if this is your problem, then I hope to have
>>> a fix very soon.
>>> Carter
>>> On Jul 9, 2009, at 11:36 AM, Phillip Deneault wrote:
>>>> Ummm, I think so? You know your code best. :-)
>>>>
>>>> If I'm reading this right, you have a pool of file handles to use. You
>>>> believe the pool is working fine except that its ignoring the srcid when
>>>> choose which entry in the pool to use? If so, yes, that is very possibly
>>>> what could be happening.
>>>>
>>>> Phil
>>>>
>>>> Carter Bullard wrote:
>>>>> OK!!! Progress.
>>>>> So, if I'm reading this right, rasplit() seems to be ignoring the srcid
>>>>> on
>>>>> some outputs?
>>>>> Rasplit() does cache the output files, and tries to be smart about open
>>>>> file descriptors, so its not reading a record, opening the output file,
>>>>> writing a single record, and closing the file. We keep a list of open
>>>>> files per srcid, when the "$srcid" keywork is in the output filename.
>>>>> It must be that we're confusing which file to use?
>>>>> Does this seem to describe where we are?
>>>>> Carter
>>>>> On Jul 9, 2009, at 11:18 AM, Phillip Deneault wrote:
>>>>>> Carter Bullard wrote:
>>>>>>> One thing to check is whether rasplit() is generating a file somewhere
>>>>>>> else
>>>>>>> in your file system, say if the "srcid" is screwy, or if time goes to
>>>>>>> zero. Are your dates in the file name looking alright?
>>>>>>
>>>>>> More progress. The time boundaries look ok, but if I compare all file
>>>>>> sizes of that time period, one is always WAY bigger than the other ones
>>>>>> and it seems to be randomly picked as to which file, in which $srcid
>>>>>> directory it will be.
>>>>>>
>>>>>>> I would recommend that you add a few more directories in your target
>>>>>>> path.
>>>>>>> Unix has a bad performance issue when the files/directory get above
>>>>>>> say
>>>>>>> 200 or so. Thats why I add a %Y/%m/%d for the slices, so that the
>>>>>>> file
>>>>>>> count doesn't get too high.
>>>>>>
>>>>>> This isn't so much of a problem since my retention on these files is
>>>>>> relatively small, only a day or two. They are being processed into
>>>>>> other things and then removed.
>>>>>>
More information about the argus
mailing list