summaryrefslogtreecommitdiffstats
path: root/config-model/src/test/java/com/yahoo/config
Commit message (Collapse)AuthorAgeFilesLines
* Warn instead of failing on validation errors in manual zonesJon Bratseth2021-06-081-2/+3
| | | | | | The purpose of validation overrides is to protect production from destructive changes by requiring a confirmation for such changes. This is unnecessary in manuall deployed environments.
* Revert "Revert "Searchdefinition/schemas cleanup""Harald Musum2021-06-052-3/+4
|
* Revert "Searchdefinition/schemas cleanup"Arnstein Ressem2021-06-042-4/+3
|
* Use generateSchemds and remove generateSearchDefinitionHarald Musum2021-06-042-3/+4
|
* searchdefinitions -> schemasHarald Musum2021-06-031-6/+6
|
* Cluster controllers are not tied to content clusters anymoreHarald Musum2021-06-011-6/+0
|
* Assign Slobroks only on cluster controllersJon Bratseth2021-05-241-14/+39
|
* Warn when specifying resource limits in hosted VespaHarald Musum2021-05-201-4/+0
| | | | | Want to throw and fail deployments, but need to make sure nobody i doing this first, so warn for now
* Merge pull request #17742 from vespa-engine/arnej/no-splunk-on-admin-nodesArne H Juul2021-05-061-0/+83
|\ | | | | skip adding logforwarder on hosts if they have membership with cluste…
| * unconditional -> include-adminArne Juul2021-05-061-1/+1
| |
| * test logforwarder not-in-adminArne Juul2021-05-061-0/+83
| |
* | Only include non-retired nodes in redundancy computationJon Bratseth2021-05-061-5/+33
|/ | | | | | | | | | Usually redundancy is down-adjusted to match the nodes per group, but not if the node count requested is high enough but downscaled by capacity policies. The redundancy is also adjusted down if there are not enough nodes in the actual allocation, which normally handles this case, but not if there are retired nodes. That is fixed in this PR by not counting retired nodes when making that adjustment.
* Merge pull request #17641 from vespa-engine/bratseth/query-profile-referencesJon Bratseth2021-04-281-8/+0
|\ | | | | Use owner's dimensions in BackedOverridableQueryProfile
| * Use owner's dimensions in BackedOverridableQueryProfileJon Bratseth2021-04-281-8/+0
| | | | | | | | | | | | | | | | Use owner's dimensions rather than the backed's in BackedObverridableQueryProfile. This matters when a profile references a profile with different but overlapping dimensions and both the owner and referred profile assigns values for the same paths, within their respective dimension spaces.
* | Add Slobroks to admin cluster controllersJon Bratseth2021-04-281-5/+19
|/
* Merge pull request #17559 from ↵Jon Marius Venstad2021-04-231-107/+24
|\ | | | | | | | | vespa-engine/bratseth/clustercontroller-cleanup-take-3 Bratseth/clustercontroller cleanup take 3
| * Revert "Merge pull request #17550 from ↵Jon Bratseth2021-04-221-107/+24
| | | | | | | | | | | | | | vespa-engine/revert-17544-bratseth/clustercontroller-cleanup-take-2" This reverts commit 8b894d8815dac50d07d4ef82ae83343915b8f8dd, reversing changes made to 7bb93176d1c15704f89d69fa6daaaefdb19692c9.
* | Make sure we are able to output context for validation errorsHarald Musum2021-04-232-5/+23
|/
* Revert "Bratseth/clustercontroller cleanup take 2"Jon Bratseth2021-04-221-24/+107
|
* Revert "Merge pull request #17534 from ↵Jon Bratseth2021-04-221-107/+24
| | | | | | | vespa-engine/revert-17528-bratseth/clustercontroller-cleanup" This reverts commit bc3caa587f166d02e90b8bb81a1656a6cdd32715, reversing changes made to 972c1f6f1131d1eb9af3f8f072131a97d1436cd5.
* Correct impossible redundancyJon Bratseth2021-04-211-0/+33
|
* Allocate retired nodes in addition to what's requestedJon Bratseth2021-04-211-24/+57
|
* Merge pull request #16818 from ↵Harald Musum2021-03-101-57/+137
|\ | | | | | | | | vespa-engine/jonmv/cluster-controller-migration-cleanup Always use dedicated cluster controllers when hosted (except in some …
| * Always use dedicated cluster controllers when hosted (except in some tests)Jon Marius Venstad2021-03-051-57/+137
| |
* | Enough to reserve 0.5Gb memory on content nodesJon Marius Venstad2021-03-031-1/+1
|/
* Apply max heap size for metrics proxy only for nodes in cluster of type 'admin'Harald Musum2021-03-011-1/+1
|
* Make dedicated cluster-controllers non-exclusiveJon Marius Venstad2021-02-251-23/+0
|
* Verify ZK cannot be set up in combined clustersJon Marius Venstad2021-02-241-0/+34
|
* Remove support for unused <controllers> under <content>Jon Marius Venstad2021-02-221-90/+1
|
* Test models with dedicated cluster controllersJon Marius Venstad2021-02-191-0/+84
|
* Remove unused testsJon Marius Venstad2021-02-191-103/+2
|
* Change default test flavour to fit cluster controllersJon Marius Venstad2021-02-191-67/+67
|
* Merge pull request #16533 from vespa-engine/bratseth/generalize-nodes-messagerJon Bratseth2021-02-161-0/+1
|\ | | | | Generalize error message and make tests more realistic
| * Generalize error message and make tests more realisticJon Bratseth2021-02-161-0/+1
| |
* | Merge pull request #16518 from ↵Henning Baldersheim2021-02-151-5/+5
|\ \ | |/ |/| | | | | vespa-engine/geirst/reserve-memory-for-content-node-processes-in-hwinfo-memory Reserve memory for other processes on the content node the proper way…
| * Reserve memory for other processes on the content node the proper way by ↵Geir Storli2021-02-151-5/+5
| | | | | | | | | | | | including it in hwinfo memory. This removes the need to scale the write filter limit used for memory.
* | Non-functional changesJon Bratseth2021-02-151-0/+2
|/
* Remove unused codeHarald Musum2021-02-091-54/+0
| | | | We will not use container nodes for cluster controllers anymore
* Remove methodHarald Musum2021-01-201-5/+0
|
* Cleanup creation of host provisionerHarald Musum2021-01-081-0/+6
|
* Print that XML is invalid according to schemaHarald Musum2021-01-072-2/+2
|
* SimplifyHarald Musum2020-12-151-32/+23
|
* Add (ignored, since it is failing) test for chaning nodes used by cluster ↵Harald Musum2020-12-151-0/+52
| | | | controllers
* Tag added servers as joiningMartin Polden2020-12-111-0/+41
|
* Allow suboptimal node count in container with ZK during retirementMartin Polden2020-12-021-0/+18
|
* Set stateful property for relevant clustersMartin Polden2020-12-011-0/+38
|
* Update ↵Jon Bratseth2020-11-121-1/+1
| | | | | config-model/src/test/java/com/yahoo/config/model/provision/ModelProvisioningTest.java Co-authored-by: Harald Musum <musum@verizonmedia.com>
* Add a testJon Bratseth2020-11-121-0/+27
|
* Support 'exclusive' node attribute for content clustersHarald Musum2020-10-161-0/+24
|
* Defer changes until restartJon Bratseth2020-10-122-2/+2
| | | | | | | | | In container clusters, if it is determined that a restart is required to swicth to a new config generation, mark that config as applicable only after a restart. This avoids wasting resources on a config change just before a restart, and the potential of prematurely applying config changes which depend on other changes which are not effective before a restart has been done.