All notable changes to this project will be documented in this file.
The format is based on Keep a Changelog and this project adheres to Semantic Versioning.
- Add additional HTTP headers for call to Themis from Convey
- Change default branch name to
main
- on connect retry, requery jwt only if it failed before
- put two timestamps in connection health file; start conn and current
- change health file update interval to 240sec
- use jitter in backoff delay
- sendMessage to check cloud status == ONLINE before sending
- when killed with SIGTERM, close will use msg in close reason file.
- Add callback handler for ping status change event
- Fixed nopoll_conn_unref crash
- Update retry timestamp in connection-health-file
- fix so that parodus can be killed, even if in a connection fail loop
- provide signal handlers so we shut down properly when INCLUDE_BREAKPAD active
- send status code and reason in websocket close message
- dont try to install handler for signal 9
- Add pause/resume heartBeatTimer
- parodus event handler to listen to interface_down and interface_up event
- Pause connection retry during interface_down event
- Update to use nanomsg v. 1.1.4
- requestNewAuthToken will clear the token if it fails.
- request auth token on every retry, not just after 403
- update to use nopoll v 1.0.2
- Fix re-registration fails that lose a socket
- Fix mutex error in service alive
- Security: Mutual Authentication (mTLS or two way TLS)
- Rename command line options for MTLS cert and Key
- Security: Added support to use auth token during initial connect to cloud
1.0.2 - 2019-02-08
- Refactored connection.c and updated corresponding unit tests
- Additional
/cloud-statusand/cloud-disconnectfields. - Switched from nanomsg (Release 1.1.2) to NNG (Release v1.0.1)
- Memory leak fixes
- Changed connection logic (connection.c) for retries, and added unit test
- Partner-id comparison made case insensitive
- Reverted from NNG to nanomag (v1.1.2)
- reverted temporary CMake reference to https://github.com/bill1600/seshat
- Added log for time difference of parodus connect time and boot time
- added NULL check for device mac id in upstream retrieve message handling
- backoff retry to include find_servers in loop (connection.c)
- backoff max is max count not max delay
- used mutex protection to make client list and nn_sends thread safe
- put mutex lock into get_global_node
- change svc alive from a thread to a function called every 30 sec from main
- shut down tasks properly
- fix memory leaks
- Fixed memory leak in upstream event message flow
- Fixed crash in CRUD request processing
- Fixed issue on RETRIEVE respone processing
- Enabled valgrind
- Fixed main loop to keep calling svc_alive_task during a cloud disconnect and retry
- change svc alive back to a separate thread. Shut it down with pthread_cond_timedwait
- Refactored Upsteam RETRIEVE flow
- Fix re-registration to call nn_shutdown and nn_close, so we don't lose a socket.
1.0.1 - 2018-07-18
- Added command line arguments for secure token read and acquire.
- DNS check for . to allow custom device handling logic (like test boxes, refurbishing, etc).
- Add check for matching
partner-idfor incoming WRP messages based on command line argument passed. - CRUD operations to local namespace is now supported.
- Via CRUD calls, the custom metadata tag values are now supported.
- Configurable security flag and port.
- Default HTTP port to 80 and HTTPS port to 443.
- Updates to how
nopollis called have been implemented.
- 64 bit pointer fixes (#144, #145).
- Handle
403error from talaria (#147). - Several additional NULL pointer checks & recovery logic blocks added.
- A scenario where ping requests are not responded to was fixed.
- Added command line arguments for secure token read and acquire. Token presented to cloud for authentication verification.
- Added support for override functions to support drop root capabilities for parodus process
1.0.0 - 2017-11-17
- Added the basic implementation.
- Initial creation