[Bioclusters] openmosix cluster again

[Hong Zhang]

So for blast, are you using blastwww?

On Tue, 6 Jan 2004, Mart=EDn Sarachu wrote:

> I think Joe is right, we are not saying contradictory things. The only pa=
tched=20
> tool we use is BLAST.
> EMBOSS, for example, is the exact same version you would use in a non-clu=
ster=20
> environment. EMBOSS jobs are quite variable in its run-times and resource=
s they=20
> need. Jobs that requires multiples accesses to a database perform very we=
ll, and=20
> jobs that do very little (check available databases) don't even migrate.
> EMBOSS apps are not paralell, so you don't get a reduction in the run-tim=
e of a=20
> single job (in fact there is some little overhead for the migration), wha=
t you=20
> get is a reduction on the time it would take to run 20 jobs.
> There's a nearly-official patch (MASK patch) for openMosix that allows to=
=20
> migrate shared-memory applications (Apache, PostgreSQL).
>=20
>=20
> Regards,
>=20
> martin
>=20
> Joseph Landman wrote:
> > Hi Hong:
> >=20
> > I don't think these are contradictory.  Martin has been able to use=20
> > OpenMosix with a variety of tools.  If you note, he used the patched=20
> > version of BLAST (which specifically disables pthreads).  The other=20
> > tools are likely also either patched, or compiled without pthreading.=
=20
> > Postgres is a database, and you really don't want it migrating from its=
=20
> > data store.  Apache is a web server, and likewise, you don't want it=20
> > migrating (extra latency etc.).
> >=20
> > Joe
> >=20
> > hong.zhang@research.dfci.harvard.edu wrote:
> >=20
> >> Dear Joe and Martin,
> >>
> >> Thanks for your instant reply. But it seems you two have contrary=20
> >> opinion.
> >> I am kinda confused. Would you please give me more explanation?
> >> I attached both of your answers below.
> >>
> >> --------------------------------------------------------------------
> >> Martin said:
> >>
> >> openMosix works great for our site (AR.EMBnet). We have many bio web=
=20
> >> tools
> >> (EMBOSS, SMS, WebPhylip) here and all of them (except SRS) work fine w=
ith
> >> openMosix. They migrate very well.
> >> On top of openMosix we use SGE for some jobs users submit by=20
> >> command-line.
> >> We use the patched version of BLAST to be able to run it (migrate) und=
er
> >> openMosix.
> >> ----------------------------------------------------------------------=
----=20
> >>
> >> Joe said
> >>
> >>>   Openmosix will not migrate pthreaded jobs.  Postgres and the web
> >>> interfaces (apache based) are all pthreaded.
> >>>
> >>>   What open mosix is as far as bio-clusters are concerned, is a tool =
to
> >>>
> >>> do a "semi-automatic" load balance.  Jobs which meet specific criteri=
a
> >>> can be migrated to less busy nodes.  For this to make sense, you woul=
d
> >>> need a) the jobs to require long running times (as compared to potent=
ial
> >>> migration time), b) sufficiently low IO and other resource utilizatio=
n
> >>> such that the migration will not generate hot-spots (resource
> >>> contention, such as network bandwidth for accessing a file over NFS).
> >>>
> >>>   It is good for some types of jobs, though in a shared nothing
> >>> environment (cluster) with limited network bandwidth and relatively h=
igh
> >>> latencies, its isn't necessarily better than the queuing systems.
> >>>
> >>> Joe
> >>>
> >>
> >>
> >>
> >>
> >> _______________________________________________
> >> Bioclusters maillist  -  Bioclusters@bioinformatics.org
> >> https://bioinformatics.org/mailman/listinfo/bioclusters
> >=20
> >=20
>=20
>=20

--=20
Hong Zhang, MIS
Bioinformatics Analyst
Dana Farber Cancer Institute
Harvard Medical School
44 Binney St, D1510A
Boston MA 02115
Email: hong.zhang@research.dfci.harvard.edu
Phone: 617-632-3824
Fax: 617-632-3351