| Commit message (Collapse) | Author | Age | Files | Lines |
| |
|
|\
| |
| |
| |
| | |
vespa-engine/geirst/correct-name-of-docs-matched-metrics
Correct the name used for *.docs_matched metrics (which are of type L…
|
| |
| |
| |
| | |
LongCountMetric).
|
|\ \
| |/
|/|
| |
| | |
vespa-engine/vekterli/export-aggregated-visit-and-deletebuckets-metrics
Export aggregated visit and deletebuckets metrics
|
| |
| |
| |
| |
| |
| |
| | |
This brings exported cross-thread aggregate metrics for persistence
threads in line with SPI metrics. Note that "visit" here really
means invocation of iterate() on the SPI, but the name remains due
to historical reasons.
|
|\ \
| |/
|/|
| |
| | |
vespa-engine/vekterli/also-log-config-changes-for-bouncer-component
Log config changes from Bouncer component
|
| |
| |
| |
| |
| | |
Also change debug output to be Slime structure, since the encoded
string does not seem to have had the intended semantics.
|
|\ \
| | |
| | | |
Nonfunctional changes
|
| | | |
|
|\ \ \
| | | |
| | | |
| | | |
| | | | |
vespa-engine/hmusum/build-only-required-model-versions-in-ap-southeast-1
Build only the minimal set of models needed in prod.ap-southeast-1
|
|/ / / |
|
|\ \ \
| | | |
| | | |
| | | |
| | | | |
vespa-engine/hakonhall/add-timeout-to-set-node-state-calls-from-orchestrator
Add timeout to set-node-state calls from Orchestrator
|
| | | | |
|
|\ \ \ \
| | | | |
| | | | |
| | | | |
| | | | | |
vespa-engine/bratseth/double-account-internal-redeploy-correctly
Propagate internalRedeploy to the RawConfig instance
|
|/ / / / |
|
|\ \ \ \
| |_|_|/
|/| | |
| | | |
| | | | |
vespa-engine/jvenstad/keep-revision-when-forcing-version-change
Keep revision target when forcing version target
|
|/ / / |
|
|\ \ \
| | | |
| | | | |
Jvenstad/deployment pipeline
|
| | | | |
|
|/ / /
| | |
| | |
| | | |
This reverts commit c6a24ec54ad6a59b5b5c602d9358372c41c314f7.
|
|\ \ \
| | | |
| | | |
| | | |
| | | | |
vespa-engine/revert-6234-jvenstad/hosted-vespa-deployment-pipeline
Revert "Jvenstad/hosted vespa deployment pipeline"
|
| | | | |
|
|\ \ \ \
| |/ / /
|/| | |
| | | |
| | | | |
vespa-engine/geirst/reduce-feeding-concurrency-setting-in-proton
Reduce default feeding.concurrency from 0.5 to 0.2.
|
| | | |
| | | |
| | | |
| | | | |
This should reduce problems with query latency spikes during re-distribution of data.
|
|\ \ \ \
| |/ / /
|/| | |
| | | |
| | | | |
vespa-engine/vekterli/debug-log-when-config-is-received-on-content-node
Debug log when new config is received on content node
|
| | | | |
|
|\ \ \ \
| | | | |
| | | | | |
translate any unknown log levels numerically (some customers chose to…
|
| | | | |
| | | | |
| | | | |
| | | | | |
com.twitter.logging.Logger which has ERROR but with a different number).
|
|\ \ \ \ \
| |/ / / /
|/| | | | |
Fix warning
|
| | | | | |
|
|\ \ \ \ \
| | | | | |
| | | | | |
| | | | | |
| | | | | | |
vespa-engine/jvenstad/hosted-vespa-deployment-pipeline
Jvenstad/hosted vespa deployment pipeline
|
| | | | | | |
|
| | | | | | |
|
| | | | | | |
|
|\ \ \ \ \ \
| | | | | | |
| | | | | | |
| | | | | | |
| | | | | | | |
vespa-engine/geirst/support-attribute-settings-for-struct-field-attributes
Support attribute settings (etc. fast-search) for struct field attrib…
|
| | | | | | | |
|
|\ \ \ \ \ \ \
| | | | | | | |
| | | | | | | | |
let last fetched summary override
|
| | | | | | | | |
|
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | | |
* instead of using the first summary that contains a field,
use the last one. This is probably more useful in the
(very few) cases where you can see the difference.
|
|\ \ \ \ \ \ \ \
| |_|_|_|_|_|_|/
|/| | | | | | |
| | | | | | | |
| | | | | | | | |
vespa-engine/mpolden/only-check-active-and-reserved
Only consider nodes in states active and reserved
|
| | | | | | | |
| | | | | | | |
| | | | | | | |
| | | | | | | | |
This matches the states checked by the internal convergence checker.
|
|\ \ \ \ \ \ \ \
| | | | | | | | |
| | | | | | | | | |
Inline macros in macros, not just first or second phase top level
|
|/ / / / / / / / |
|
|\ \ \ \ \ \ \ \
| | | | | | | | |
| | | | | | | | | |
Reuse allocation decisions
|
| | | | | | | | |
| | | | | | | | |
| | | | | | | | |
| | | | | | | | |
| | | | | | | | |
| | | | | | | | |
| | | | | | | | |
| | | | | | | | |
| | | | | | | | |
| | | | | | | | |
| | | | | | | | |
| | | | | | | | |
| | | | | | | | |
| | | | | | | | | |
During deployment we first prepare the application on one config server
then notifies the other config servers (by setting state=PREPARED)
that they should load it. That will cause allocation requests against
the node repo to be done once over on the those config servers.
These simultaneous requests against the node repo may cause ut
to time out waiting for the node repo's application lock if
these requests take a somewhat long time.
This change reads existing allocations instead of redoing
allocation requests against the node repo. The existing allocation
is saved (on the first config server) before state=PREPARED is set.
|
|\ \ \ \ \ \ \ \ \
| | | | | | | | | |
| | | | | | | | | | |
Nonfunctional changes only
|
| |/ / / / / / / / |
|
|\ \ \ \ \ \ \ \ \
| |_|_|_|/ / / / /
|/| | | | | | | | |
Fix hash calculation for directory file references
|
| | | | | | | | |
| | | | | | | | |
| | | | | | | | |
| | | | | | | | | |
Use the path to each file when walking through directory
|
|\ \ \ \ \ \ \ \ \
| |_|_|_|/ / / / /
|/| | | | | | | |
| | | | | | | | |
| | | | | | | | | |
vespa-engine/hmusum/run-tenants-maintainer-in-all-zones
Run TenantsMaintainer in all zones
|