[9.0] more on RAM matching#8414
Conversation
fstagni
force-pushed
the
90_more_RAM_matching
branch
8 times, most recently
from
7801c52 to
76938f4
Compare
fstagni
force-pushed
the
90_more_RAM_matching
branch
2 times, most recently
from
24d38b8 to
7e9bf4a
Compare
| except IndexError: # we are in Jenkins | ||
| J.setInputSandbox([find_all("exe-script-with-input.py", "/home/dirac", "DIRAC/tests/Workflow")[0]]) | ||
| J.setExecutable("exe-script-with-input.py", "", "helloWorld.log") | ||
| J.setInputData(["/dteam/user/f/fstagni/test/testInputFile.txt"]) |
There was a problem hiding this comment.
Out of curiosity, where do this file and testInputFileSingleLocation.txt come from?
Do we have access to them?
There was a problem hiding this comment.
I simply uploaded myself...
What do you mean with "access"? If another user can access it? I believe yes.
| resourceDict[name] = resourceDescription[name] | ||
|
|
||
| for name in multiValueMatchFields: | ||
| for name in singleValueDefFields + multiValueMatchFields + ("MaxRAM",): |
There was a problem hiding this comment.
Any reason for not having a rangeValueMatchFields in TaskQueueDB that would contain MaxRAM?
So that we avoid having hardcoded value like MaxRAM spread throughout the code.
There was a problem hiding this comment.
rangeValueMatchFields also includes MinRAM, which here is not needed.
| nProcessors = int(nProcessors) | ||
| except ValueError: | ||
| nProcessors = None | ||
| for param, key, limit, increment in [(maxRAM, "MB", 1024 * 1024, 256), (nProcessors, "Processors", 1024, 1)]: |
There was a problem hiding this comment.
I'm curious, don't you need that mechanism anymore?
There was a problem hiding this comment.
Update: I think I get it, it's because you now define range values, right?
There was a problem hiding this comment.
The mechanism is still there, just it's only needed for Processors (since RAM requirements are not stored anymore as tags).
| if not retVal["OK"]: | ||
| # If table doesn't exist (e.g., old installation), log a warning but continue | ||
| # This provides backward compatibility | ||
| if "doesn't exist" in retVal["Message"] or "Table" in retVal["Message"]: |
There was a problem hiding this comment.
That sounds a bit fragile but I guess you don't have much choice.
There was a problem hiding this comment.
It's indeed a bit fragile, but already the chances of hitting this case are close to 0, as the table SHOULD be existing here (I am tempted to actually remove this check...).
fstagni
force-pushed
the
90_more_RAM_matching
branch
from
7e9bf4a to
d3c8663
Compare
4 participants
BEGINRELEASENOTES
*WMS
NEW: added a new table to TaskQueueDB for RAM requirements and matching
ENDRELEASENOTES
The new table should be automatically created when the DB class is instantiated (at service restart)