Merge lp:~jpds/launchpad/mirrorprober-next-probe-schema into lp:launchpad/db-devel
| Status: | Work in progress |
|---|---|
| Proposed branch: | lp:~jpds/launchpad/mirrorprober-next-probe-schema |
| Merge into: | lp:launchpad/db-devel |
| Diff against target: |
93 lines (+36/-20) 4 files modified
database/sampledata/current-dev.sql (+10/-10) database/sampledata/current.sql (+10/-10) database/schema/comments.sql (+1/-0) database/schema/patch-2207-50-0.sql (+15/-0) |
| To merge this branch: | bzr merge lp:~jpds/launchpad/mirrorprober-next-probe-schema |
| Related bugs: |
| Reviewer | Review Type | Date Requested | Status |
|---|---|---|---|
| Björn Tillenius (community) | db | 2010-04-15 | Approve on 2010-04-15 |
| Stuart Bishop | db | 2010-04-15 | Approve on 2010-04-15 |
|
Review via email:
|
|||
Commit Message
Added next_probe timestamp column to distributionmirror to enable future manual mirror probes.
Description of the Change
= Summary =
Added a next_probe timestamp column to distributionmirror, this will allow us to manually request probes of mirrors as requested on bug #196175.
This branch adds the needed schema, models to follow.
- 9251. By Jonathan Davies on 2010-04-15
-
Make next_probe default to now, as we have already approved mirrors.
| Stuart Bishop (stub) wrote : | # |
Fine here. The timestamp is overkill for the specified use case, but may be useful for other things like spreading the checks over the whole day rather than in one big batch.
We could drop the enabled flag if we made the new column nullable, but I think keeping it is clearer and we are not dealing with enough data to worry about efficiency.
It needs a comment for the new column added to database/
I notice we still have an unused Mirror table. I guess that should be dropped. Time for me to run some checks and take out the garbage.
| Stuart Bishop (stub) wrote : | # |
Allocated patch number is patch-2207-50-0.sql
| Björn Tillenius (bjornt) wrote : | # |
Looks good, since it's consistent with checkwatches and the branch puller:
< BjornT> jpds: did you discuss the implementation with anyone? i'm not.
sure exactly what you're planning to do.
< jpds> BjornT: Change the distribution mirror to be more like the branch.
puller so that we can have manual probes.
< BjornT> jpds: can you point me to the relevant part of the branch.
puller?
< jpds> BjornT: next_mirror for branches, and next_check for bug watches.
- 9252. By Jonathan Davies on 2010-04-15
-
Renamed patch as per stub's allocation.
- 9253. By Jonathan Davies on 2010-04-15
-
Added comment for DistributionMir
ror.next_ probe. - 9254. By Jonathan Davies on 2010-04-15
-
Change next_probe so that it can be null (don't probe mirrors which this value).
- 9255. By Jonathan Davies on 2010-04-15
-
Added checks and constraints as per stub's suggestions.
| Stuart Bishop (stub) wrote : | # |
With ec2 land now.
| Stuart Bishop (stub) wrote : | # |
This db patch will need to be landed in tandem with the code changes.
The distribution-
- 9256. By Jonathan Davies on 2010-04-16
-
Amended patch-2207-50-0.sql so that it only sets next_probe to now for enabled
and official mirrors. - 9257. By Jonathan Davies on 2010-04-16
-
Updated sampledata for patch-2207-50-0.sql changes.
Unmerged revisions
- 9257. By Jonathan Davies on 2010-04-16
-
Updated sampledata for patch-2207-50-0.sql changes.
- 9256. By Jonathan Davies on 2010-04-16
-
Amended patch-2207-50-0.sql so that it only sets next_probe to now for enabled
and official mirrors. - 9255. By Jonathan Davies on 2010-04-15
-
Added checks and constraints as per stub's suggestions.
- 9254. By Jonathan Davies on 2010-04-15
-
Change next_probe so that it can be null (don't probe mirrors which this value).
- 9253. By Jonathan Davies on 2010-04-15
-
Added comment for DistributionMir
ror.next_ probe. - 9252. By Jonathan Davies on 2010-04-15
-
Renamed patch as per stub's allocation.
- 9251. By Jonathan Davies on 2010-04-15
-
Make next_probe default to now, as we have already approved mirrors.
- 9250. By Jonathan Davies on 2010-04-15
-
Added database patch patch-2207-49-0.sql - adds a new column to
distributionmirror for bug #196175.

Fine here. The timestamp is overkill for the specified use case, but may be useful for other things like spreading the checks over the whole day rather than in one big batch.
We could drop the enabled flag if we made the new column nullable, but I think keeping it is clearer and we are not dealing with enough data to worry about efficiency.
It needs a comment for the new column added to database/ schema/ comments. sql
I notice we still have an unused Mirror table. I guess that should be dropped. Time for me to run some checks and take out the garbage.