5.11.0
Improvements
- meta2: Vacuum with synchronous resync when database is small (6e2e581)
- Python API: read whole EC fragments from rawx (e76f977)
- CLI: replace GreenletProfiler by yappi (0a7e51b)
- Python API: ramp up read buffer size (3236495)
- account: Avoid listing shards when listing buckets (18525d4)
- CI: Improve the copyright check script (b11a262, cba8fe4)
- meta2: Add request ID in "container.new" and "container.state" events (eedb9b4)
- meta1: Add request ID in "account.services" event (bdb83ff)
- rawx: Send "chunk.new" event when linking chunks (670e350)
- rawx: Update Go Modules (a3b80e2)
- CI: Make some tests really fast (efae32b)
- sqliterepo: make election debug messages clearer (89be978)
- sqliterepo: on corruption, display the name of the base (67f6bad)
Bug fixes
- sqliterepo: Extract user-agent to pass it in meta2 events (23d9306)
- server,sqliterepo: fix request max_size (867dfc5)
- server: protect ASN.1 server against HTTP requests (f64272e)
- rawx(Go): fix computation hash path (29726a7)
- sqliterepo: detect Zookeeper >= 3.6.0 (b2e92bf)
- account: fix refresh bucket lua script (288533f)
Warning
If, in an installed configuration, the variable sqliterepo.dump.max_size
is defined, it will absolutely be necessary to add the variable server.request.max_size
, paying attention to its chosen value (cf. https://github.com/open-io/oio-sds/blob/5.x/Variables.md#sqliterepodumpmax_size).