summaryrefslogtreecommitdiffstats
path: root/node-repository
Commit message (Collapse)AuthorAgeFilesLines
* Use architecture in more places in node repoHarald Musum2022-03-036-8/+13
|
* Ignore architecture when adding or subtracting NodeResourcesHarald Musum2022-03-024-12/+18
| | | | | Need to solve this long-term, but ignore for now as provisoning will not work as this is now.
* Support architecture in node repositoryHarald Musum2022-02-283-6/+34
| | | | | * Persist architecture for nodes * Use architecture when talking to node repo
* Use flag source ref from node repoJon Bratseth2022-02-271-3/+3
|
* Reuse node indexesJon Bratseth2022-02-255-13/+64
|
* Use enum as parameter typeHarald Musum2022-02-252-13/+17
|
* Add architeecture to node repo rest apiHarald Musum2022-02-2442-126/+202
|
* Update testsHarald Musum2022-02-2319-29/+29
|
* Merge pull request #21331 from vespa-engine/bratseth/out-of-capacityJon Marius Venstad2022-02-239-53/+54
|\ | | | | Replace 'out of capacity' by 'node allocation failure'
| * Replace 'out of capacity' by 'node allocation failure'Jon Bratseth2022-02-239-53/+54
| |
* | Always park deprovisioning infrastructure nodeMartin Polden2022-02-232-9/+11
| | | | | | | | | | | | PR #20879 assumes that all deprovisioning child nodes should be parked, but this is not the case for infrastructure as it breaks assumptions done by dynamic provisioning.
* | Revert "Retry deprovisioning if child is un-retired"Martin Polden2022-02-233-47/+18
|/
* Remove deprovisioning child node when readiedMartin Polden2022-02-221-1/+7
|
* Retry deprovisioning if child is un-retiredMartin Polden2022-02-213-17/+40
|
* Do not offer node moves to reserved hostsMartin Polden2022-02-111-0/+2
|
* Merge pull request #21141 from vespa-engine/mpolden/update-aclMartin Polden2022-02-112-2/+3
|\ | | | | Stop trusting port 4080 on proxy nodes
| * Stop trusting port 4080 on proxy nodesMartin Polden2022-02-112-2/+3
| |
* | Park hosts with high fail countMartin Polden2022-02-111-3/+11
|/
* Do not allocate nodes to suspended hostsValerij Fredriksen2022-02-034-17/+51
|
* Add Orchestrator to NodeRepositoryValerij Fredriksen2022-02-0316-51/+38
|
* Make Nodes::canAllocateTenantNodeTo non-staticValerij Fredriksen2022-02-034-14/+16
|
* Fix potential NPEValerij Fredriksen2022-02-031-1/+1
|
* Merge pull request #21006 from vespa-engine/mpolden/consider-newly-activeJon Bratseth2022-02-015-76/+104
|\ | | | | Configure load balancer with nodes that will become active
| * Configure load balancer with nodes that will become activeMartin Polden2022-02-015-76/+104
| |
* | Close the futureJon Bratseth2022-01-311-5/+8
| |
* | Add method for issuing a query and waiting for it to be executedHarald Musum2022-01-311-2/+10
| | | | | | | | | | Needs to be done for some queries (e.g. 'alter table) from Questdb version 6.1.3 and later.
* | Upgrade questdbHarald Musum2022-01-281-1/+1
|/ | | | Upgrade to a version that supports JDK 17
* Merge pull request #20938 from vespa-engine/bratseth/modular-profilesJon Bratseth2022-01-261-1/+1
|\ | | | | Bratseth/modular profiles
| * No functional changesJon Bratseth2022-01-251-1/+1
| |
* | Increase down grace time while nodes are suspendedJon Bratseth2022-01-253-30/+44
| |
* | No functional changesJon Bratseth2022-01-251-48/+50
|/
* Move retired nodes to dirty if wantToDeprovisionValerij Fredriksen2022-01-206-8/+9
|
* Extend ThreadedHttpReqHndlr instead of deprecated LoggingReqHndlrgjoranv2022-01-172-5/+5
|
* Merge pull request #20820 from vespa-engine/hakonhall/remove-dev-systemMartin Polden2022-01-177-44/+0
|\ | | | | Remove dev system
| * Remove dev systemHåkon Hallingstad2022-01-147-44/+0
| |
* | Replace use of Guava's UncheckedTimeoutException with our ownBjørn Christian Seime2022-01-142-2/+2
|/
* Increase warmup durationJon Bratseth2022-01-123-3/+3
| | | | Warming up a container takes a long time.
* Read only relevant states when state filter is presentMartin Polden2022-01-125-71/+92
|
* Merge pull request #20775 from vespa-engine/mpolden/read-onceHarald Musum2022-01-121-5/+5
|\ | | | | Avoid repeatedly reading nodes for every known node type
| * Avoid repeatedly reading nodes for every known node typeMartin Polden2022-01-121-5/+5
| |
* | Bump cache sizeMartin Polden2022-01-121-4/+4
|/ | | | | Makes all nodes fit in the cache, even in our largest zones. Today excessive eviction happens because nodes are constantly shuffled in and out of the cache.
* Merge pull request #19819 from vespa-engine/olaa/remove-old-clusterid-tagBjørn Meland2022-01-111-2/+0
|\ | | | | Remove old clusterid tag
| * Remove old clusterid tagOla Aunrønning2021-11-011-2/+0
| |
* | Don't report 0 success on collisions if we ignore themJon Bratseth2022-01-111-1/+2
| |
* | Merge pull request #20684 from vespa-engine/freva/node-idValerij Fredriksen2022-01-072-3/+4
|\ \ | | | | | | Use UUID for Linux container nodes
| * | Use UUID for Linux container nodesValerij Fredriksen2022-01-062-3/+4
| | |
* | | unify java warningsArne H Juul2022-01-061-8/+0
|/ / | | | | | | | | * these were stricter than in parent, but to simplify we can just use compiler args from parent
* | Add DeployLogger field to DomConfigPayloadBuildergjoranv2022-01-041-1/+1
| |
* | Add InfrastructureApplication in application-modelHåkon Hallingstad2022-01-031-4/+4
| |
* | Add flag for deactivating routing for an applicationMartin Polden2021-12-162-3/+29
| |