1 |
On 06/12/2016 04:01 PM, Andrew Savchenko wrote: |
2 |
> On Thu, 2 Jun 2016 10:42:03 -0500 james wrote: |
3 |
>> On 06/01/2016 06:20 PM, Justin Bronder wrote: |
4 |
>> > Due to a lack of time and the fact I don't use any of these packages |
5 |
>> > anymore, they are all up for grabs. |
6 |
>> > |
7 |
>> > - media-gfx/openmesh [no project] |
8 |
>> > - sys-cluster/ganglia [cluster] |
9 |
>> > - sys-cluster/ganglia-web [cluster] |
10 |
>> > - sys-cluster/torque [cluster] |
11 |
>> > - sys-cluster/munge [cluster] dependency of sys-cluster/torque |
12 |
>> > - sys-cluster/mpe2 [cluster] |
13 |
>> > |
14 |
>> > Also, if there's anyone out there using the science overlay and empi |
15 |
>> > who's feeling motivated, that work still needs a champion to get it |
16 |
>> > into the main tree. If not, I'll probably drop it in a few months |
17 |
>> > and open openmpi and mpich2 to project maintenance as well. I |
18 |
>> > haven't been involved in HPC for over a decade now, it's time to |
19 |
>> > pass the torch. |
20 |
>> |
21 |
>> |
22 |
>> Hello Justin, |
23 |
>> |
24 |
>> I've been working on cluster ebuilds for a while (Apache Mesos, spark, |
25 |
>> etc). I'm willing to proxy maintain these except torque. Assuming there |
26 |
>> are no users of torque on gentoo (bgo seems inactive...it's dead; how |
27 |
>> would I know?). |
28 |
> |
29 |
> Hey, don't be too quick on your judgement :) We are using torque |
30 |
> (and maui) on three HPC setups, though we had to patch torque for |
31 |
> various features. |
32 |
|
33 |
|
34 |
Ian picked up torque:: |
35 |
|
36 |
"I can step up and maintain or co-maintain sys-cluster/torque , I use |
37 |
it at work and have contributed to it in the past." |
38 |
|
39 |
Which was written on 6/2/16 by axs@g.o. |
40 |
|
41 |
|
42 |
> By the way, what scheduler/resource manager are you using for HPC? |
43 |
|
44 |
I'm more focused on small clusters and the ability to 'reboot' identical |
45 |
hardware into a variety of HPC or container cluster configurations. My |
46 |
immediate goal is fast and small clusters to run 'cluster-benchmarks' |
47 |
and compare different cluster offerings, for a particular problem. The |
48 |
emphasis is the hardware which is identical, so the cluster codes are |
49 |
the only difference. NO elaborate Schedulers or Framework schedulers, |
50 |
just testing everything, atm. |
51 |
|
52 |
That said, you may want to look at Mesos, as that cluster OS, it |
53 |
supports a myriad of Schedulers and Framework schedulers. It fact it |
54 |
encourages custom scheduler development. |
55 |
|
56 |
http://mesos.apache.org/ |
57 |
|
58 |
http://mesos.apache.org/documentation/latest/app-framework-development-guide/ |
59 |
|
60 |
|
61 |
|
62 |
> slurm ebuild seems to be even in worse state than torque one (2.6.3 |
63 |
> is present, 16.05.0 is the latest upstream). |
64 |
|
65 |
I do not know why more gentoo devs have not been bitten by the |
66 |
sys-cluster bug. Slurm is still widely used. SchedMD is hiring. |
67 |
|
68 |
|
69 |
> Best regards, |
70 |
> Andrew Savchenko |
71 |
|
72 |
|
73 |
> Best regards, |
74 |
> Andrew Savchenko |
75 |
|
76 |
|
77 |
James |
78 |
|
79 |
|
80 |
[1] |
81 |
http://mesos.apache.org/documentation/latest/app-framework-development-guide/ |