Crunch1-in-Crunch2 » History » Version 1
Tom Clegg, 06/11/2015 07:36 PM
1 | 1 | Tom Clegg | {{>TOC}} |
---|---|---|---|
2 | |||
3 | h1. Crunch1-in-Crunch2 (DRAFT) |
||
4 | |||
5 | Detail about how Crunch2 runs jobs that were written for Crunch1. |
||
6 | |||
7 | See |
||
8 | * Crunch2 [[Jobs API]] |
||
9 | |||
10 | h2. Background |
||
11 | |||
12 | In order for Crunch2 to replace Crunch1, Crunch2 must be capable of running jobs that rely on Crunch1's API. Examples: |
||
13 | * run-command |
||
14 | * arv-run (via run-command) |
||
15 | * existing tutorial/example jobs |
||
16 | * user scripts based on existing tutorials |
||
17 | |||
18 | h2. Requirements |
||
19 | |||
20 | Crunch1 jobs rely on the following pieces: |
||
21 | * arv-run-pipeline-instance fulfills a pipeline instance by submitting and monitoring jobs using the Crunch1 API. (This piece could be replaced by a Crunch2 equivalent if necessary; however, users may have written job-submission/monitoring scripts along the same lines and it's ideal if these also continue to work.) |
||
22 | * Keep mount available within the container |
||
23 | * Some environment variables (CRUNCH_SRC, ARVADOS_API_*, etc) |
||
24 | * jobs and job_tasks APIs |