[Biocorba-l] Re: cvs.bioperl.org running ensembl perl corba

[Jason Stajich]

----- Original Message -----
From: "Ewan Birney" <birney@ebi.ac.uk>
To: "Brad Chapman" <chapmanb@arches.uga.edu>
Cc: "Jason Stajich" <jason@chg.mc.duke.edu>
Sent: Monday, April 16, 2001 4:03 AM
Subject: Re: cvs.bioperl.org running ensembl perl corba


>
>
> Great work guys. I think we should move some of this disucssion onto
> either ensembl-dev or biocorba (or both?)
>
done.  not sending to ensembl-dev just yet, again was afraid server would
get swamped considering number of individuals on that list who are
interested in different access methods to ensembl data.

For those we are bringing this into discussion late:  I  have setup an
ensembl server and ensembl biocorba implementation on top of it running on
cvs.bioperl.org.   Brad is testing it with his biopython test suite.  I'm
not quite sure the server is ready for public testing.  One issue is what
type of bandwith requirements are we going to require - we don't want to
wear out our welcome at the Genetics Institute.  Fetching whole chromosome
annotations can be large (well actually this doesn't work with current
ensembl code -- could be my using live trunk with ens 0.8.0 data)

Some other issues we are having - versions of ORBit > 0.5.3 (Brad - remind
us?)  are not playing nicely with CORBA::ORBit perl module.  I know that
ORBit 0.5.7 is not behaving with freebsd machine so we're not clear what the
problem is.  If this is a full blown server will need to run ORBit-mt.  Not
sure what the reprecussions will be with CORBA::ORBit perl packge.  Chris
Mungall has had a fair amount of experience here and may want to weigh in
when we can specific reproducable problems.

> i guess one quesiton is - what will happen if we put up a corba server off
> sanger centre - (a) is it feasible to draw things across the internet and
> (b) will our server die from over use.
>
> I think there is only one way to find out.
>
This was my intention of setting it up on cvs.bioperl.org to get a feel for
the server load and network load before trying to open it up as a public
interface.  There are some issues with the mysql engine on the openbsd
server - I'm getting crapouts on certain large queries.  Not sure what is
exact problem as I haven't really debugged.

Brad - reason code is still complaing about Obj cvs DBAdaptor code is the
ensembl code on cvs.bioperl is from the anonymous repository and it wasn't
updated immedietly after I checked into ensembl - will do a cvs update this
morning and restart server -- suspect that will set things right.


>
> -----------------------------------------------------------------
> Ewan Birney. Mobile: +44 (0)7970 151230, Work: +44 1223 494420
> <birney@ebi.ac.uk>.
> -----------------------------------------------------------------
>
>