Currently, when shutting down the server (via SIGINT or SIGTERM), the
shutdown closes any open connections after only 10 seconds (via a
context.WithTimeout). This does not provie a lot of time for longer
operations, such as listing versions, or downloading a larger module zip
file.
When running in Kubernetes, and scaling instances or changing config,
this causes a lot of dropped connections and gateway errors. 10 seconds
is arguably much too short, and should be configurable.
This commit increases that default to 60 seconds, and adds a config
variable to allow users to specify their desired timeout.
* read redis lock options from config to support custom TTL & timeout
* fix test
* fix typo
* downgrade to bsm/redislock@v0.7.2 to prevent usage of beta go-redis version
* revert test changes
* return error for invalid lock config
* update config parsing test
* udpate docs to include redis lock config
* fix test
* set default max retries to 10
* reduce default redis lock timeout to 15s
* update default TTL to 15mins
Co-authored-by: Manu Gupta <manugupt1@gmail.com>
* Reverting PR 1650
https://github.com/gomods/athens/pull/1650 was a big change to the
authentication/authorization code, which we have decided to pull
out and potentially move into a separate process/project
Signed-off-by: Aaron Schlesinger <aaron@ecomaz.net>
* removing commented, unused code
Signed-off-by: Aaron Schlesinger <aaron@ecomaz.net>
* removing more commented, unused code
Signed-off-by: Aaron Schlesinger <aaron@ecomaz.net>
* removing more unused code
Signed-off-by: Aaron Schlesinger <aaron@ecomaz.net>
* add config options for redis password
* redis password test and failure test
* changed redis name and made a minimal redis.conf
add drone volume
volume fix
remove volumes from drone.
repoint redis path
Co-authored-by: Aaron Schlesinger <70865+arschles@users.noreply.github.com>
Co-authored-by: Marwan Sulaiman <marwan.sameer@gmail.com>
* Fixing the build failures from a missing GCS Key
Accidentally introduced in https://github.com/gomods/athens/pull/1428
* Adding a TODO to the drone build
* adding a step to test the Athens server startup
* Adding the JSONKey back to the default config
* use temporary image
* Switching back to the old env var for the GCS JSON Key
cc/ @marwan-at-work
* removing the ServiceAccount. we want to stick with JSONKey
* reverting to original key fetching code
* fixing build err
* bumping chart ver
* Add GCS to helm chart
* base64 encode serviceAccount in the chart
* Standardize GCP Service Account references
* pipe storage.gcp.serviceAccount to quote
* cmd/proxy/actions: pass NoSumPatterns to GoGetFetcher
* add gosum unit tests
* upgrade appveyor to 1.13
* fix test typo
* check appveyor version
* AppVeyor has no 1.13
* use logger
This CL addresses issue #1177 introducing two new configuration params
allowing to activate pprof.
- `ATHENS_ENABLE_PPROF`
- `ATHENS_PPROF_PORT`
pprof won't be exposed by default.
* actions: support proxying sum db urls
* remove proxy prefix
* add docs for checksum db
* more docs
* typo
* typo
* typo
* typo
* typo
* typo
* typo
* typo
* move checksum db into its own section
* add azblob sf stasher
* use http status const
* rm line
* cleanup, comments
* add test
* fix test
* skip test it no account key provided
* fix stash
* introduce stash timeout
* Rename config.dev.toml to config-example.toml
* Updating helm docs to point to the new helm chart repository (#1102)
* Updating helm docs to point to the new helm chart repository
Fixes https://github.com/gomods/athens/issues/1099
Follow-up to https://github.com/gomods/athens/pull/1097/files
* fixes
* Helm chart: Fix ingress setup (#1086)
* Helm chart: Fix ingress setup
The ingress part of the Helm chart contained references to
not-existing templates and values. Additionally, the deployment used
some no longer available health-check URLs which prevented it from
being marked as ready.
* Make readiness and liveness paths depend on the image.tag
* Add read from .athens.toml (from pwd, then home)
* Remove config lookup from HOME
* Revert "Rename config.dev.toml to config-example.toml"
This reverts commit 9c02956406.
* Fix comment
* Allow mgo.ParseURL to set [/database]
As per Michael's solution we need to be able to set the database from
either the parsed connection string, or fallback to athens if not set.
Signed-off-by: Chris M <me@christophermills.co.uk>
* Remove TimeoutDuration which was removed #928
Signed-off-by: Chris M <me@christophermills.co.uk>
* Correct conf_test to accommodate for changes
I'd missed out the addition of DefaultDBName from the test, this has now
been corrected
Signed-off-by: Chris M <me@christophermills.co.uk>
* Make sure gofmt runs automatically on this machine
Y'know simple stuff.
Signed-off-by: Chris M <me@christophermills.co.uk>
* Added new cataloger interface
* Implementing catalog protocol
* Propagated to protocol and over
* First round of fixes
* S3 almost ready, need to be tested
* Going on with testing s3
* Better testing with s3
* Simplified catalog tests
* Preparing gcp tests to access a gcp instance
* Fixing initialization errors
* Removed some prints
* Gcp ready, to be tested
* Gcp working
* Aligned bucket mock to catalog method
* Switched res payload to json
* Added catalog method to all storage instances
* Added catalog method to unsupported storages
* Fixed with pool test
* Restored tests
* Fixed gcp constructor
* Implemented catalog for fs
* Removed trace
* E2e tests, fixed fs
* Fixed module name return value
* Added cataloger method to azure storage
* Added docs
* Changed pagesize parameter name
* Fixed gofmt error
* Added json tags to result. Fixed lint warning
* Removed extra line
* Changed not implemented error to http.KindNotImplemented
* Checking for inequality on results
* Lower-cased json keys
* Added cleaning of path separator
* Fixed review comments
* Add catalog endpoint for mongo
* Add omitempty to ID
* Fix catalog tests
* update for next token
* fix e2e
* Make query readable
* Fix language and e2e script
* remove new line
