[2.2] MAAS should confirm the success of DNS reloads by checking the zone serial
Bug #1710308 reported by
Mike Pontillo
This bug affects 1 person
Affects | Status | Importance | Assigned to | Milestone | |
---|---|---|---|---|---|
MAAS |
Fix Released
|
Critical
|
Blake Rouse | ||
2.2 |
Fix Released
|
Critical
|
Unassigned |
Bug Description
MAAS currently reloads the named configuration after every DNS change. When a large number of changes occur, this has negative side-effects for bind9. (For context, see bug #1710278.)
MAAS should wait for confirmation that the reload is complete before starting another update.
See also: bug #1712205 (throttling of DNS reload requests) and bug #1710310 (better service monitoring, separate from the reload confirmation itself).
Related branches
~blake-rouse/maas:fix-1710308-2.2
- Blake Rouse (community): Approve
-
Diff: 332 lines (+196/-5)4 files modifiedsrc/maasserver/dns/config.py (+6/-0)
src/maasserver/dns/tests/test_config.py (+19/-0)
src/maasserver/region_controller.py (+46/-3)
src/maasserver/tests/test_region_controller.py (+125/-2)
~blake-rouse/maas:improved-dns-reload
Merged
into
maas:master
- Mike Pontillo (community): Approve
-
Diff: 332 lines (+196/-5)4 files modifiedsrc/maasserver/dns/config.py (+6/-0)
src/maasserver/dns/tests/test_config.py (+19/-0)
src/maasserver/region_controller.py (+46/-3)
src/maasserver/tests/test_region_controller.py (+125/-2)
Changed in maas: | |
importance: | High → Critical |
Changed in maas: | |
status: | Triaged → In Progress |
assignee: | nobody → Blake Rouse (blake-rouse) |
Changed in maas: | |
status: | In Progress → Fix Committed |
Changed in maas: | |
milestone: | 2.3.0 → 2.3.0alpha2 |
summary: |
- [2.2] MAAS should throttle bind reloads + [2.2] MAAS should confirm the success of DNS reloads by checking the + zone serial |
description: | updated |
description: | updated |
description: | updated |
Changed in maas: | |
status: | Fix Committed → Fix Released |
To post a comment you must log in.
You might find the logic of the "watershed" CLI tool useful. Wrapping the "rndc reload" call with the watershed command might also work around the underlying reload hanging issue. It's in the chiark-utils-bin package.