summaryrefslogtreecommitdiffstats
path: root/node-repository
Commit message (Collapse)AuthorAgeFilesLines
* Support restricting on host OS when allocating new nodesHåkon Hallingstad2021-01-263-4/+30
|
* Use the same activation transaction to write LBsValerij Fredriksen2021-01-211-6/+9
|
* Ensure transitive dependencies of vespajlib are present on test classpathBjørn Christian Seime2021-01-201-0/+6
| | | | | | Some modules misuse container-dev as a compile/test dependency or have container-test missing as test dependency. Depend on container-test where it does not create a cycle. Otherwise depend directly on vespajlib without altering scope of existing transitive vespajlib dependency.
* Suggest only when usefulJon Bratseth2021-01-1911-47/+82
| | | | | | | | - Closes a loophole where the suggestion made will be lower than the current allocation if the current allocation is the max need observed over the last week. - Since we now store the suggestion even if it is current we check at read time whether to suggest, and then also refrain from making suggestions inside the autoscaling interval.
* Merge pull request #16081 from vespa-engine/bratseth/suggest-on-human-scaleValerij Fredriksen2021-01-187-19/+100
|\ | | | | Bratseth/suggest on human scale
| * Set suggestion to highest oberved over a weekJon Bratseth2021-01-182-1/+25
| | | | | | | | | | Since suggestions are consumed by humans they should change on the time scale of human decision making.
| * Add timestamp to suggestionsJon Bratseth2021-01-157-19/+76
| |
* | Merge pull request #16078 from vespa-engine/mpolden/ignore-testerBjørn Meland2021-01-181-0/+1
|\ \ | | | | | | Ignore tester application in allocation metrics
| * | Ignore tester application in allocation metricsMartin Polden2021-01-181-0/+1
| |/
* | adding nonactive to total for fractionBjorn Meland2021-01-162-2/+2
| |
* | Allocate max half of hosts actually available in zoneJon Bratseth2021-01-152-6/+11
| |
* | No functional changesJon Bratseth2021-01-154-12/+10
|/
* Switch from mem.total.util to mem.util in autoscalingJon Bratseth2021-01-142-3/+3
| | | | | mem.total.util included disk cache which makes it unsuitable as a regulation target.
* Merge pull request #15966 from ↵Martin Polden2021-01-1211-87/+205
|\ | | | | | | | | vespa-engine/hakonhall/ensure-fresh-node-with-lock Acquire locks for children when patching
| * Minor fixes after review roundHåkon Hallingstad2021-01-123-5/+12
| |
| * Acquire lock for children when patchingHåkon Hallingstad2021-01-114-38/+79
| |
| * Rename to lockAndGetHåkon Hallingstad2021-01-089-20/+20
| |
| * Ensure fresh node with lockHåkon Hallingstad2021-01-0811-65/+135
| |
* | Revert "Revert "Count lock timeout as unsuccessful run""Martin Polden2021-01-081-2/+2
| |
* | Revert "Count lock timeout as unsuccessful run"Harald Musum2021-01-081-2/+2
| |
* | Remove special handling of lock timeoutMartin Polden2021-01-081-1/+1
| |
* | Count lock timeout as unsuccessful run for exclusive maintainersMartin Polden2021-01-071-2/+2
|/
* Merge pull request #15941 from vespa-engine/mpolden/exclusive-switch-metricJon Bratseth2021-01-074-57/+112
|\ | | | | Report exclusive switch metric per cluster
| * Report exclusive switch metric per clusterMartin Polden2021-01-074-57/+112
| |
* | Revert "Revert "Make clients use orchestratorStatus instead of allowedToBeDown""Håkon Hallingstad2021-01-061-1/+1
|/
* Revert "Make clients use orchestratorStatus instead of allowedToBeDown"Håkon Hallingstad2021-01-061-1/+1
|
* Merge pull request #15911 from ↵Håkon Hallingstad2021-01-061-1/+1
|\ | | | | | | | | vespa-engine/hakonhall/make-clients-use-orchestratorstatus-instead-of-allowedtobedown Make clients use orchestratorStatus instead of allowedToBeDown
| * Make clients use orchestratorStatus instead of allowedToBeDownHåkon Hallingstad2021-01-051-1/+1
| |
* | Revert "Revert "Reapply "Upgrade to Curator 4"""Harald Musum2021-01-051-0/+1
| |
* | Merge pull request #15915 from vespa-engine/mpolden/rebalancing-fixJon Bratseth2021-01-054-34/+100
|\ \ | | | | | | Avoid moving nodes already on exclusive switch
| * | Avoid moving nodes already on exclusive switchMartin Polden2021-01-053-27/+100
| | |
| * | Remove stdout spamMartin Polden2021-01-051-7/+0
| | |
* | | Revert "Reapply "Upgrade to Curator 4""Harald Musum2021-01-051-1/+0
| |/ |/|
* | Merge branch 'master' into revert-14062-revert-14057-hmusum/upgrade-to-curator-4Harald Musum2021-01-0161-573/+1427
|\ \
| * | Don't repeat targetJon Bratseth2020-12-171-1/+1
| | |
| * | Track current resource changes in autoscaling statusJon Bratseth2020-12-172-16/+19
| |/
| * Require minimum measurements depending on scaling windowJon Bratseth2020-12-103-15/+20
| |
| * Do more checksJon Bratseth2020-12-101-10/+25
| |
| * Compare just numbersJon Bratseth2020-12-101-1/+1
| |
| * Rename close to awaitShutdownJon Marius Venstad2020-12-103-4/+4
| |
| * Split (async) shutdown and (sync) wait for it, and use lists in all ↵Jon Marius Venstad2020-12-092-72/+43
| | | | | | | | maintainer owners
| * Fix message used with verifyTargetMartin Polden2020-12-081-1/+1
| |
| * Log allocation details of node being movedMartin Polden2020-12-081-1/+1
| | | | | | | | | | ...and not for the host we're moving from (which is always allocated to `hosted-vespa.tenant-host`)
| * Improve status messagesJon Bratseth2020-12-071-2/+2
| |
| * Merge pull request #15682 from vespa-engine/mpolden/detect-clashing-namesJon Marius Venstad2020-12-072-3/+70
| |\ | | | | | | Reject endpoints with clashing non-compactable IDs
| | * Reject load balancers with clashing non-compactable IDsMartin Polden2020-12-042-3/+70
| | |
| * | Start a little more optimisticJon Bratseth2020-12-041-1/+1
| | |
| * | Adaptive scaling windowJon Bratseth2020-12-0410-61/+145
| | |
| * | Merge pull request #15624 from vespa-engine/mpolden/cleanupMartin Polden2020-12-042-7/+7
| |\ \ | | |/ | |/| Remove redundant compatibility with old format
| | * Update node-repository testsMartin Polden2020-12-032-7/+7
| | |