Gentoo Archives: gentoo-cluster

From: Hanni Ali <hanni.ali@×××××.com>
To: gentoo-cluster@l.g.o
Subject: Re: [gentoo-cluster] cluster or distributed queue, general question
Date: Fri, 11 Jan 2008 12:02:05
Message-Id: 789d27b10801110402rbe36429k29576e81828df863@mail.gmail.com
In Reply to: RE: [gentoo-cluster] cluster or distributed queue, general question by Jos Houtman
1 You mentioned DrQueue, that batch processing app fulfills many of your
2 requirements. I wrote an ebuild for it too which is on bugs.gentoo
3
4 Hanni
5
6 On 11/01/2008, Jos Houtman <jos@×××××.nl> wrote:
7 >
8 > Thanx for the replies,
9 >
10 > I will certainly try the beowulf mailinglist next week,
11 > The suggestions are great gearman seems to fit except for being perl
12 > oriented (atleast that's what I read sofar).
13 > Torque/pbs, don't know about that yet.
14 >
15 > Thanx again,
16 >
17 > Jos Houtman
18 > System administrator Hyves.nl
19 > email: jos@×××××.nl
20 >
21 > -----Original Message-----
22 > From: Robin H. Johnson [mailto:robbat2@g.o]
23 > Sent: donderdag 10 januari 2008 15:30
24 > To: gentoo-cluster@l.g.o
25 > Subject: Re: [gentoo-cluster] cluster or distributed queue, general
26 > question
27 >
28 > On Thu, Jan 10, 2008 at 02:59:27PM +0100, Jos Houtman wrote:
29 > > For my master thesis I took up a project that requires mapping of a
30 > number of statically defined parallel jobs into a more dynamic
31 > environment that allows better scaling.
32 > > The situation as described below let me to believe a cluster or
33 > distributed queue (DrQueue?) solution is necessary. For the situation
34 > see [situation] at the end of this email.
35 > Off the top of my head, many of your requirements are available in two
36 > totally different apps:
37 > - Gearman, written by Brad Fitzpatrick @ LiveJournal. Perl mainly, I
38 > think there are other interfaces as well to it.
39 > - Torque/PBS - somewhat less of a fit, I'm not certain about running
40 > perpetual jobs.
41 >
42 > You may also need some degree of STONITH for the job running only once
43 > during node failure case. (Say the job manager crashes, the job is still
44 > running, but you have no control of it. You need to zap it hard).
45 >
46 > --
47 > Robin Hugh Johnson
48 > Gentoo Linux Developer & Infra Guy
49 > E-Mail : robbat2@g.o
50 > GnuPG FP : 11AC BA4F 4778 E3F6 E4ED F38E B27B 944E 3488 4E85
51 > --
52 > gentoo-cluster@l.g.o mailing list
53 >
54 >
55
56
57 --
58 E-mail: hanni.ali@×××××.com
59 Mobile: +44 (0) 7985580147
60 My Blog: http://ainkaboot.co.uk/blogs/hanni/
61 Website: http://ainkaboot.co.uk http://drqueue.org