Add least processing time distribution algorithm#1
Conversation
simon-weber
left a comment
simon-weber
left a comment
There was a problem hiding this comment.
looking good! left a minor comment about the error handling
| "'--lpt-data <lpt-data-filepath>' to be specified as " | ||
| "well. Falling back to hash-ring algorithm." | ||
| )) | ||
| self.algorithm = self.ALGORITHM_HASH_RING |
There was a problem hiding this comment.
thoughts on just crashing here+below instead of falling back? my guess is it's intended to defend against things higher up in the stack (eg failing to download a data file), which does seem like what we want but might not be the best place to handle it
There was a problem hiding this comment.
right, good point. lemme give that some thought.
There was a problem hiding this comment.
actually, yeah, falling back is definitely bad in the case of the failed download since that might have happened at only one node and then the different nodes might successfully execute different algorithms.
good call. i'll revise this to just crash and fail for these cases.
There was a problem hiding this comment.
ah, yeah, good point. fwiw, so long as nose doesn't write out a nosetests.xml our builds will consider it unstable, so we could hook up retries at that level if needed.
plboerboom
force-pushed
the
lpt_distribution
branch
3 times, most recently
from
25dad61 to
afbd403
Compare
|
hey @simon-weber -- made some error handling updates based on your comments. can you take another look? |
plboerboom
force-pushed
the
lpt_distribution
branch
from
fb33d1e to
988ebc4
Compare
plboerboom
force-pushed
the
lpt_distribution
branch
2 times, most recently
from
cc8617a to
8ee3ccc
Compare
plboerboom
force-pushed
the
lpt_distribution
branch
from
8ee3ccc to
4c40b49
Compare
|
lgtm 👍 |
2 participants
No description provided.