summaryrefslogtreecommitdiff
Commit message (Collapse)AuthorAgeFilesLines
...
* Add method to ensure a valid topic name (#1238)Niklas Mollenhauer2017-10-072-3/+53
|
* Fix typoJeff Widman2017-10-061-1/+1
|
* Explicitly check for `None` rather than FalseJeff Widman2017-10-061-1/+1
| | | | | If the group leader somehow gets in a state that it has an empty partition assignment, then `self._assignment_snapshot` will be `{}` which evaluates to `False`. So `self._subscription.mark_for_reassignment()` will never be triggered, even if `self._assignment_snapshot != self._metadata_snapshot`. Fixes the symptoms of https://github.com/dpkp/kafka-python/issues/1237 although I suspect there's an additional bug in that case that triggers the condition of the the group leader getting an empty partition assignment.
* Fix Fetcher.PartitionRecords to handle fetch_offset in the middle of ↵Dana Powers2017-10-052-3/+31
| | | | compressed messageset (#1239)
* Fix grammarJeff Widman2017-10-041-1/+1
|
* Small fixes to SASL documentation and logging; validate security_protocol ↵Dana Powers2017-10-031-21/+26
| | | | (#1231)
* Remove raw traceback (#1221)dbgasaway2017-10-011-3/+5
|
* Merge pull request #1213 from ↵Taras Voinarovskyi2017-09-135-13/+13
|\ | | | | | | | | dpkp/1212-stop-using-mutable-types-for-default-arg-values Stop using mutable types for default arg values
| * Stop using mutable types for default arg values1212-stop-using-mutable-types-for-default-arg-valuesJeff Widman2017-09-125-13/+13
| | | | | | | | | | | | | | | | Using mutable types for default args is typically a no-no unless their surprising behavior is being explicitly abused, for an explanation see: http://effbot.org/zone/default-values.htm Fix #1212
* | Fix typoJeff Widman2017-09-121-1/+1
|/
* Merge pull request #1200 from buptljy/offsetTaras Voinarovskyi2017-09-102-14/+0
|\ | | | | remove beginning/end offsets request version limit
| * remove beginning/end offsets request version limitLiao Jiayi2017-09-042-14/+0
|/
* Remove a few unused imports (#1188)James Lamb2017-08-294-4/+3
| | | | | * Removed a few unused imports * Added note on socketpair monkey-path
* BrokerConnection receive bytes pipe (#1032)Dana Powers2017-08-154-93/+121
|
* Drop unused sleep kwarg to poll (#1177)Dana Powers2017-08-155-31/+25
|
* Update sphinx/RTD changelog for 1.3.4Dana Powers2017-08-132-1/+68
|
* bump version for developmentDana Powers2017-08-131-1/+1
|
* Update compatibility docs re: 0.11 brokers and python 3.6Dana Powers2017-08-133-13/+13
|
* Release 1.3.41.3.4Dana Powers2017-08-132-1/+59
|
* Add 0.11.0.0 to travis test matrix, remove 0.10.1.1; use scala 2.11 ↵Dana Powers2017-08-132-30/+18
| | | | artifacts (#1176)
* Initialize metadata_snapshot in group coordinator (#1174)Dana Powers2017-08-131-6/+9
|
* Select on sockets to avoid busy polling during bootstrap (#1175)Dana Powers2017-08-131-0/+2
|
* Merge pull request #1161 from dpkp/issue1036_offset_by_timeTaras Voinarovskyi2017-08-077-68/+609
|\ | | | | Added support for offsets_for_times, beginning_offsets and end_offsets APIs.
| * Added unit tests for fetcher's `_reset_offset` and related functions.Taras Voinarovskiy2017-08-073-7/+199
| |
| * Added `beginning_offsets` and `end_offsets` API's and fixed @jeffwidman ↵Taras Voinarovskiy2017-08-074-17/+142
| | | | | | | | review issues
| * Fix test for older brokersTaras Voinarovskiy2017-08-071-1/+1
| |
| * Changed retrieve_offsets to allow fetching multiple offsets at onceTaras Voinarovskiy2017-08-073-100/+174
| |
| * Fix test for older brokersTaras Voinarovskiy2017-08-071-2/+4
| |
| * Added basic support for offsets_for_times API. Still needs to group by nodes ↵Taras Voinarovskiy2017-08-076-21/+169
|/ | | | and send in parallel.
* Add private map of api key -> min/max versions to BrokerConnection (#1169)Dana Powers2017-08-062-11/+58
|
* Fixed Issue 1033.Raise AssertionError when decompression unsupported. (#1159)webber2017-08-051-0/+7
|
* conn: Catch ssl.EOFErrors on Python3.3 so we close the failing conn (#1162)Hannu Valtonen2017-08-011-1/+3
|
* Merge pull request #1160 from dpkp/issue1155Taras Voinarovskyi2017-07-311-0/+1
|\ | | | | Make UnknownTopicOrPartitionError retriable error
| * Make UnknownTopicOrPartitionError retriable errorissue1155Taras Voinarovskiy2017-07-301-0/+1
|/
* Fix batch expiry messages to state secondsDominic Evans2017-07-261-3/+3
| | | | | time.time() is seconds since epoch, and the deltas are also calculated in seconds
* Use for join-time-max and sync-time-max metrics Max() measure function (#1146)Alexey Pervushin2017-07-201-2/+2
|
* added gssapi support (Kerberos) for SASL (#1152)Harald2017-07-201-2/+75
|
* Add note, that `max_in_flight_requests_per_connection>1` may change order or ↵Taras Voinarovskyi2017-07-171-1/+4
| | | | messages (#1149)
* producer: fix produce timeout message (#1151)Mika Eloranta2017-07-171-1/+1
|
* Protocol updates for 0.11.0.0 (#1127)Dana Powers2017-07-097-41/+508
|
* Use logging's built-in string interpolationJeff Widman2017-07-072-6/+5
|
* change_subscription called only when necessary (#1132)Petr Šebek2017-07-071-2/+3
| | | | | | When we are using subscription by pattern change subscription is called every metadata update even when nothing changes. This PR ensures that change subscription is called only when set of topics changes.
* Fix typoJeff Widman2017-07-011-1/+1
|
* producer: Set exit timeout to 0 for atexit handler to match __del__ (#1126)Hannu Valtonen2017-06-191-1/+1
| | | | | | | | Hit a problem with pytest hitting the atexit handler and waiting for close() timeout forever at teardown. This commit makes atexit close() equivalent to __del__ behavior, namely using timeout of 0 for close() completion. If you need a longer timeout you should be setting it explicitly.
* KIP-144: Exponential backoff for broker reconnections (#1124)Dana Powers2017-06-194-22/+69
|
* Backoff on unavailable group coordinator retry (#1125)Dana Powers2017-06-191-0/+2
|
* Fixup for #1085 -- only check for changed metadata on disconnected nodesDana Powers2017-06-181-17/+23
|
* Deal with brokers that disappear, reappear with different IP address (#1085)Mike Fischer2017-06-181-1/+12
| | | | | | | | | | | | | | | | | | | | When KafkaClient connects to a broker in _maybe_connect, it inserts into self._conns a BrokerConnection configured with the current host/port for that node. The BrokerConnection remains there forever, though, so if the broker's IP or host ever changes, KafkaClient has no way to deal with this. The fix is to compare the latest metadata with the current node's connection, and if the host/IP has changed, decommission the old connection and allow a new one to be created. There's also a common race condition on broker startup where the initial metadata request sometimes returns an empty list of brokers, but subsequent requests behave normally. So, we must deal with broker being None here. This change is conservative in that it doesn't remove the connection from self._conns unless the new broker metadata contains an entry for that same node with a new IP/port.
* Do not test fetched_size with fetch_max_bytes since 0.11 brokers will return ↵Dana Powers2017-06-181-3/+0
| | | | more data than 0.10
* Increase max_buffer_size for test_large_messagesDana Powers2017-06-181-2/+8
|