[1/3] kudu git commit: KUDU-2297 (part 3): refactor process-wide stack collection out of /stacks

2018-02-23 Thread dralves
Repository: kudu Updated Branches: refs/heads/master 416b3018a -> 00815045f KUDU-2297 (part 3): refactor process-wide stack collection out of /stacks Previously a bunch of logic to collect all the stacks from the process was in the /stacks path handler. This logic is relatively generic and

[2/3] kudu git commit: KUDU-2297 (part 4): periodically dump stacks to diagnostics log

2018-02-23 Thread dralves
KUDU-2297 (part 4): periodically dump stacks to diagnostics log This modifies the diagnostics log to periodically dump stack traces. This is slightly complicated by the fact that symbolized stack traces can be relatively large. So, we separate the logging of symbols and stack traces. When an

[3/3] kudu git commit: Don't perform compactions when clean time has not been advanced

2018-02-23 Thread dralves
Don't perform compactions when clean time has not been advanced Due to KUDU-2233 we might not advance safe time, and thus clean time, at bootstrap causing possible corruption or crashes. This patch adds a check to make sure that clean time has been advanced at all before performing a compaction.

[1/2] kudu git commit: KUDU-2291 (part 8): fix a TSAN issue with libunwind initialization

2018-02-23 Thread danburkert
Repository: kudu Updated Branches: refs/heads/master 2a5a12fbd -> 57fe0c3db KUDU-2291 (part 8): fix a TSAN issue with libunwind initialization libunwind uses double-checked locking for initialization, which isn't technically safe. We previously tried to work around this by calling into the

[2/2] kudu git commit: KUDU-2191 (5/n): Add Kerberos SASL support to the HMS client

2018-02-23 Thread danburkert
KUDU-2191 (5/n): Add Kerberos SASL support to the HMS client The bulk of this commit is adding a new Thrift transport type, SaslClientTransport, which facilitates SASL GSSAPI negotiation, as well as integrity/privacy channel protection. The new transport is based on Impala's version with some

[2/2] kudu git commit: [Spark] Simplifiy creating range partitioned tables

2018-02-23 Thread granthenke
[Spark] Simplifiy creating range partitioned tables In order to create a range partitioned table a user needs to create PartialRows representing the lower and upper bounds. In the Spark API the Kudu Schema isn’t made available which could make matching the expected schema tricky. This small

kudu git commit: fs: clarify error message at startup

2018-02-23 Thread awong
Repository: kudu Updated Branches: refs/heads/master e94556ef7 -> 2a5a12fbd fs: clarify error message at startup Currently the error messages you get when you start a Kudu instance without specifying all the expected directories is: I0205 09:26:39.961467 25499 server_base.cc:229] Could not