DNAseq: 3700 crash due to comments text

Bernd Weisshaar weisshaa at mpiz-koeln.mpg.de
Fri May 25 04:11:23 EST 2001

Dear DNAseq'ers,

since some time we have the problem that our 3700 crashes during data extra=
ction. All data (as far as we know) from the run in question are lost, and =
a complete database purge is required to get the instrument back to work. I=
nitially, we thought that the crashes were unpredictable (as did the Abi se=

Now, we have traced the problem down to our comment text in the "Comment" c=
olumn of the plt file. We use this field to transfer information from our L=
IMS to the sequence trace (!) file for subsequent automatic analysis of the=
 data, and the length did increase over time when more information was tran=
sferred in an easy-parsable (and human-readable) way. An example of an info=
rmative plt file is shown here:


This file is really informative because the data extractor (Collection-Soft=
ware 1.1; Sequencing-Analysis 3.6) did work until probe E09, but failed wit=
h F09. The error message is 'orbix not enough memory'. The difference betwe=
en E09 and F09 is three characters more in the comments of F09.

Has anybody else seen this problem? Is there any hint in the documentation =
about a length limt of the comments field (we did not find such hints)? The=
 Abi service was in this special case unable to give any further informatio=
n. Has anyone found a trick to circumvent the problem?

Obviously, we can try to limit the length of the text. But since the inform=
ation transfer via the trace file itself is very convinient in terms of dat=
a integrity, and because human readability is also useful (I do not want to=
 compress the info in a way like "chmod 771" on UNIX), I would like to have=
 the bug removed to allow more characters in the comments fields.

Thanks is advance, Bernd

PD Dr. Bernd Weisshaar
MPI fuer Zuechtungsforschung (english: MPI for Plant Breeding Research)
Carl-von-Linn=E9-Weg 10
50829 Koeln

Tel. office: +49-221-5062 850
Tel. lab: +49-221-5062 860
Fax: +49-221-5062 851
mailto:weisshaa at mpiz-koeln.mpg.de
WWW-URL: http://www.mpiz-koeln.mpg.de/~weisshaa/BW-info.html


More information about the Autoseq mailing list