[1/3] incubator-kudu git commit: Doxygen for C++ client API

Repository: incubator-kudu Updated Branches: refs/heads/master 3dd0ad2aa -> 3cb390e38 http://git-wip-us.apache.org/repos/asf/incubator-kudu/blob/3cb390e3/src/kudu/util/CMakeLists.txt -- diff --git

[3/3] incubator-kudu git commit: Doxygen for C++ client API

Doxygen for C++ client API If doxygen is available, build 'doxygen' taget to generate Doxygen docs from client.h and other Kudu C++ client API files, After the target is built, open ${CMAKE_CURRENT_BINARY_DIR}/docs/doxygen/client_api/html/index.html in your favorite browser to see the generated

[2/3] incubator-kudu git commit: Doxygen for C++ client API

http://git-wip-us.apache.org/repos/asf/incubator-kudu/blob/3cb390e3/src/kudu/client/client.h -- diff --git a/src/kudu/client/client.h b/src/kudu/client/client.h index 710768c..31fb630 100644 --- a/src/kudu/client/client.h +++

[1/2] kudu git commit: cfile-test: some test micro-optimization to avoid timeouts

Repository: kudu Updated Branches: refs/heads/master 0348499e3 -> acf093c18 cfile-test: some test micro-optimization to avoid timeouts It turns out that SCOPED_TRACE is relatively slow. Removing some SCOPED_TRACE in hot parts of the test speeds some of the test cases up by almost 2x in ASAN

[1/3] kudu git commit: Fix to KUDU-1642 patch

Repository: kudu Updated Branches: refs/heads/master 86714152d -> 8993716d7 Fix to KUDU-1642 patch This is a quick fix to the MakePredicate code. It had an unnecessary Schema argument that can be dropped by making it non-static. Change-Id: I5120368476991bdcdbf58df72931bb72d5116ab8

[2/3] kudu git commit: Update Impala docs for Impala 2.8 release

Update Impala docs for Impala 2.8 release * No longer needs to document the special 'IMPALA_KUDU' build. * Fixed syntax for new style * Added back CTAS, which has syntax (we'd mistakenly thought it was removed) * Fixed a few typos/issues elsewhere (eg use of 'int32' instead of 'int' type) *

[3/3] kudu git commit: Bump Squeasel version

Bump Squeasel version Squeasel has recently landed some security-related fixes: * OpenSSL is now dynamically linked, instead of dlopened. * Squeasel can now skip OpenSSL initialization, which prevents conflicts with Kudu's initialization. Change-Id: If5c69b412ff8f9e4c5eb1f35f86ee62e0348bf0c

[2/2] kudu git commit: [security] Simplify MakeSignature and VerifySignature impls

[security] Simplify MakeSignature and VerifySignature impls Change-Id: I5b396de67b0de8302c515f149c174061ef5fb620 Reviewed-on: http://gerrit.cloudera.org:8080/5880 Reviewed-by: Todd Lipcon Tested-by: Kudu Jenkins Project: http://git-wip-us.apache.org/repos/asf/kudu/repo Commit:

[1/2] kudu git commit: [security] adjust TLS certificate verification

Repository: kudu Updated Branches: refs/heads/master e4a649c20 -> 85fd90996 [security] adjust TLS certificate verification This commit adjusts when TLS certificate verification is performed in the client and server, and adds a test of negotiating GSSAPI with TLS. In the client, the server's

kudu git commit: [security] add channel binding to krpc

Repository: kudu Updated Branches: refs/heads/master 848180654 -> 6fec9dd52 [security] add channel binding to krpc Channel binding prevents a MITM attack when using unauthenticated TLS with Kerberos. The channel binding codepath is exercised by the existing TLS + GSSAPI negotiation test, but

[1/2] kudu git commit: tls_socket: support zero-length writes

Repository: kudu Updated Branches: refs/heads/master 8993716d7 -> 67fcfccd4 tls_socket: support zero-length writes In some cases, we try to call Write() with a zero-length buffer. This is allowed with the normal write() or writev() APIs, but OpenSSL uses a '0' return value to indicate certain

[2/2] kudu git commit: Add security-test dependency on security

Add security-test dependency on security Change-Id: I341de3a46af75ed9716eb81f22b0ef7e92029338 Reviewed-on: http://gerrit.cloudera.org:8080/5862 Reviewed-by: Todd Lipcon Tested-by: Kudu Jenkins Project: http://git-wip-us.apache.org/repos/asf/kudu/repo Commit:

[2/2] kudu git commit: [thirdparty] fix build-thirdparty.sh for bash 3.x

[thirdparty] fix build-thirdparty.sh for bash 3.x The fall-through directive ';&' works only for bash >= 4.0. This change makes it work for bash 3.x as well. Change-Id: I624e5d13b8a981491a65bae590be1f1e432710a9 Reviewed-on: http://gerrit.cloudera.org:8080/5832 Tested-by: Kudu Jenkins

kudu git commit: Rename SaslAuth to SaslMechanism

Repository: kudu Updated Branches: refs/heads/master aa90fc02c -> d47c13872 Rename SaslAuth to SaslMechanism As part of the recent RPC negotiation changes the role of this message changed slightly, and the new name is more indicative of its purpose. Change-Id:

kudu git commit: [java client] KUDU-1643 Prune hash partitions based on IN-list predicates

Repository: kudu Updated Branches: refs/heads/master e7e65954b -> 4f0677f84 [java client] KUDU-1643 Prune hash partitions based on IN-list predicates PartitionPruner is updated to search all combinations of in-list column values and prune hash partitions where possible. This also fixes a

[3/3] kudu git commit: Break cyclic dependency of token.proto on krpc

Break cyclic dependency of token.proto on krpc Change-Id: I28245c71276dd9c33c325f820310143c2872b3cb Reviewed-on: http://gerrit.cloudera.org:8080/5942 Tested-by: Kudu Jenkins Reviewed-by: Todd Lipcon Project: http://git-wip-us.apache.org/repos/asf/kudu/repo Commit:

kudu git commit: client: trust master's cert, adopt authn token

Repository: kudu Updated Branches: refs/heads/master 93df70337 -> c4c770a45 client: trust master's cert, adopt authn token This changes the client to actually act on the security-related data coming back from the ConnectToMaster() RPC. The client now adds the master's CA cert to its trusted

[1/3] kudu git commit: [security] method to check if X509 cert matches key

Repository: kudu Updated Branches: refs/heads/master c4c770a45 -> 959a7835d [security] method to check if X509 cert matches key Added a method to security::Cert interface to check whether X509 certificate matches the specified private key. Change-Id: I6732f8a1ad9ed1b1ab26b94b582af50e6b6d24b1

[2/3] kudu git commit: [security] use Cert::CheckKeyMatch() for sanity checks

[security] use Cert::CheckKeyMatch() for sanity checks Use the Cert::CheckKeyMatch() method in appropriate places. Change-Id: I47d9134377ad51af153accb00ce3ac4cf864cda1 Reviewed-on: http://gerrit.cloudera.org:8080/5938 Reviewed-by: Dan Burkert Tested-by: Kudu Jenkins

[1/7] kudu git commit: java: fix ability to connect to a real Kerberized cluster

Repository: kudu Updated Branches: refs/heads/master 959a7835d -> b7b418b5e java: fix ability to connect to a real Kerberized cluster This fixes a couple issues seen when I tried to run the Java client from within Impala against a Kerberized Kudu cluster: * Previously, the ServerInfo class

[7/7] kudu git commit: Remove sync_point.[h,cc]

Remove sync_point.[h,cc] It no longer appears to be used, and it was causing a compiler warning: /Applications/Xcode.app/Contents/Developer/Toolchains/XcodeDefault.xctoolchain/usr/bin/ranlib: file: lib/libkudu_util_exported.a(sync_point.cc.o) has no symbols Change-Id:

[5/7] kudu git commit: java: allow eclipse checkstyle to work

java: allow eclipse checkstyle to work Eclipse's checkstyle plugin doesn't properly handle paths relative to the checkstyle configuration XML file. The workaround is that it does provide a "samedir" property which can be used to construct the proper absolute path. Of course, Maven doesn't provide

[3/7] kudu git commit: java: use a netty frame decoder instead of replaying decoder

java: use a netty frame decoder instead of replaying decoder All of our inbound packets are length-prefixed. Netty provides a nice class for handing length-prefixed messages. This avoids a bunch of more custom logic we were doing for length checking, etc. Change-Id:

[6/7] kudu git commit: Rename SecureRpcHelper to Negotiator

Rename SecureRpcHelper to Negotiator Change-Id: Ia5790a8dbf1a147fd5e6bbc793bc3707456121d2 Reviewed-on: http://gerrit.cloudera.org:8080/5944 Reviewed-by: Dan Burkert Reviewed-by: Jean-Daniel Cryans Tested-by: Kudu Jenkins Project:

[2/7] kudu git commit: java: remove unused parts of SecureRpcHelper

java: remove unused parts of SecureRpcHelper This removes the unused wrap/unwrap code from SecureRpcHelper. I also removed an unnecessary function in TabletClient. Change-Id: I3b70e0319060e3f2d9cc9a0ed23da7fe51b174df Reviewed-on: http://gerrit.cloudera.org:8080/5925 Reviewed-by: Dan Burkert

[4/7] kudu git commit: Make SecureRpcHelper a Netty pipeline stage

Make SecureRpcHelper a Netty pipeline stage This decouples SecureRpcHelper from TabletClient. When negotiation is complete, it sends an "event" upstream to the TabletClient and removes itself from the pipeline. I'm hoping this decoupling will make it easier to test SecureRpcHelper in isolation

[3/3] kudu git commit: Fix TSAN segfault on Centos 6

Fix TSAN segfault on Centos 6 The OpenSSL version shipped with Centos 6 appears to segfault if a nullptr is passed to OBJ_find_sigid_algs with TSAN. I haven't been able to actually figure out why, I haven't been succesful in finding the Centos 6 source for OpenSSL, and the version it's based on

[2/3] kudu git commit: master_rpc: pass back more details from ConnectToCluster

master_rpc: pass back more details from ConnectToCluster This plumbs back the ConnectToMasterResponsePB from the ConnectToCluster RPC. This brings us one step closer to inserting the CA cert into the TlsContext and passing the token somewhere useful. Along the way, I switched to std::function

[1/2] kudu git commit: TLS-negotiation [3/n]: rename negotiation protobuf messages

Repository: kudu Updated Branches: refs/heads/master ef57bda2c -> 9d0421e80 TLS-negotiation [3/n]: rename negotiation protobuf messages SASL negotiation in the RPC layer has evolved to encompass more than just SASL. Today it includes RpcFeatureFlag negotiation, and in the future it may

kudu git commit: Change links for old releases to Apache archive

Repository: kudu Updated Branches: refs/heads/gh-pages 95b9a7abc -> 46c3cd53b Change links for old releases to Apache archive Apache doesn't like to keep old versions around on their mirror network, etc. So, change the links to use archive.apache.org instead in preparation for removing them.

[1/2] kudu git commit: TLS-negotiation [6/n]: Refactor RPC negotiation

Repository: kudu Updated Branches: refs/heads/master b9aa5dd31 -> dc8525358 http://git-wip-us.apache.org/repos/asf/kudu/blob/dc852535/src/kudu/rpc/reactor.cc -- diff --git a/src/kudu/rpc/reactor.cc b/src/kudu/rpc/reactor.cc

[2/2] kudu git commit: TLS-negotiation [6/n]: Refactor RPC negotiation

TLS-negotiation [6/n]: Refactor RPC negotiation The KRPC negotiation process now encompasses more than just SASL authentication. Feature flags are negotiated, and soon TLS encryption will be negotiated as well. In anticipation of TLS negotiation, this commit refactors the SaslClient and

kudu git commit: KUDU-1852. KuduTableAlterer should not crash with nullptr arguments

Repository: kudu Updated Branches: refs/heads/master dc8525358 -> 8a2ea8bdd KUDU-1852. KuduTableAlterer should not crash with nullptr arguments Change-Id: Ifa7c34b476111ecd33d4fd3ef5cc363a410ad76a Reviewed-on: http://gerrit.cloudera.org:8080/5797 Reviewed-by: Dan Burkert

kudu git commit: Remove some unused class members

Repository: kudu Updated Branches: refs/heads/master e367b4e0b -> b9aa5dd31 Remove some unused class members Change-Id: Ia8c7265aa1a690cc0987e634c3292f00b67f9a74 Reviewed-on: http://gerrit.cloudera.org:8080/5782 Reviewed-by: Dan Burkert Tested-by: Kudu Jenkins

kudu git commit: [security] add TLS protocol and cipher suite to negotiation trace

Repository: kudu Updated Branches: refs/heads/master 8067304f6 -> 93ec241a9 [security] add TLS protocol and cipher suite to negotiation trace No tests, but I manually verified the output looks good: 0216 14:14:45.214356 (+48us) server_negotiation.cc:484] Negotiated TLSv1.2 with cipher

[1/2] kudu git commit: [client, server]_negotiation: replace gscoped_ptr with unique_ptr

Repository: kudu Updated Branches: refs/heads/master fea449ace -> 758cf8b8d [client, server]_negotiation: replace gscoped_ptr with unique_ptr Change-Id: Ibe0151aeb54af4f54b7dbb8e019127329cdd526b Reviewed-on: http://gerrit.cloudera.org:8080/6136 Reviewed-by: Alexey Serbin

[1/2] kudu git commit: exactly_once_rpc-test: properly propagate sleep parameter

Repository: kudu Updated Branches: refs/heads/master 0430ea630 -> 60aedaaf4 exactly_once_rpc-test: properly propagate sleep parameter Change-Id: I043647661b1b1c83479b3c9b2f2a93bc3fc7d546 Reviewed-on: http://gerrit.cloudera.org:8080/5883 Tested-by: Kudu Jenkins Reviewed-by: Todd Lipcon

[2/2] kudu git commit: [rpc] Assume existence of TlsContext during negotiation

[rpc] Assume existence of TlsContext during negotiation This changes the client and server RPC negotiatiators to always assume the existence of a TlsContext for the messenger, and check whether there are valid certs in the context in order to decide whether to use TLS encryption. This will

[2/2] kudu git commit: java: implement Channel Bindings

rowables.propagate(e); +} + +chan.getPipeline().addFirst("tls", handler); sendSaslInitiate(chan); } @@ -362,6 +366,7 @@ public class Negotiator extends SimpleChannelUpstreamHandler { } ByteString data = ByteString.copyFrom(bufs); if (sslHandshakeFuture.isDon

[2/2] kudu git commit: TLS-negotiation [2/n]: remove SASL ANONYMOUS mechanism

TLS-negotiation [2/n]: remove SASL ANONYMOUS mechanism ANONYMOUS has not been enabled on any released server version going back all the way to 1.0. Adding it back an enabling it would be a backwards compatible change, so there is no harm in removing the currently unused code. Change-Id:

[1/5] kudu git commit: KUDU-1904 Don't seek on empty RLE blocks

Repository: kudu Updated Branches: refs/heads/master 3ae7fe1c9 -> 9591957b8 KUDU-1904 Don't seek on empty RLE blocks Seeking in RLE blocks enforces that the seek position is less than the number of elements in the block. If the number of elements is 0, as it is when the entire set of cells is

[2/5] kudu git commit: KUDU-1880 Prevent eager ignoring of NULLs

KUDU-1880 Prevent eager ignoring of NULLs During a nullable block scan, if a NULL set of cells is reached before the decoder evaluation is disabled (e.g. if the first rows in the block are NULL), the rows will be prematurely marked as not in the results set. If evaluating an IsNull predicate,

[4/5] kudu git commit: Make reduce flake from KUDU-1642 tests in client-test

Make reduce flake from KUDU-1642 tests in client-test This ups the timeouts so the tests should be less flaky. It also removes a couple stray zeroes and improves the flush behavior. Change-Id: Iba01eb7e9976f2fae4061ba918457c946d98c067 Reviewed-on: http://gerrit.cloudera.org:8080/5882 Tested-by:

[3/5] kudu git commit: [client-test] Reduce flakyness of TestWriteWithDeadTabletServer

[client-test] Reduce flakyness of TestWriteWithDeadTabletServer This test fails often due to an unexpected reason for the write timeout. This is related to KUDU-1466. I just commented out the assertion and left a TODO to reenable once we address the jira. Also added some info to the jira itself.

[5/5] kudu git commit: server_negotiation: improve error handling

server_negotiation: improve error handling Unlike in the client negotiator, the server negotiator should send an error response back to the client on authentication failure. We were missing this in a few cases, notably around token and cert authentication failures. This also adds a new

kudu git commit: [java-client] Add ScanToken.stringifySerializedToken

Repository: kudu Updated Branches: refs/heads/master 1f660ee9f -> 4fd0572ee [java-client] Add ScanToken.stringifySerializedToken stringifySerializedToken takes a serialized scan token, and returns a String suitable for debug printing. The string contains information sufficient to determine

svn commit: r15688 - /release/kudu/KEYS

Author: danburkert Date: Mon Sep 26 18:45:06 2016 New Revision: 15688 Log: Add code signing key for Dan Burkert Modified: release/kudu/KEYS Modified: release/kudu/KEYS == --- release/kudu/KEYS (original) +++ release

[5/6] kudu git commit: KuduPredicate::Clone on IN list predicate causes segfault

KuduPredicate::Clone on IN list predicate causes segfault Jordan Birdsell uncovered this issue where a cloned IN list predicate causes a segfault when applied to a scanner. During the clone, the values list was being 'preallocated' using the vector(int) constructor, which adds a bunch of default

[3/6] kudu git commit: build-and-test.sh: update gcovr location

build-and-test.sh: update gcovr location Coverage builds were failing because of this. Change-Id: I8874d298b6176b4deb45749847032e5c2be55955 Reviewed-on: http://gerrit.cloudera.org:8080/4567 Reviewed-by: David Ribeiro Alves Tested-by: Adar Dembo Project:

[1/2] kudu git commit: KUDU-1652 (part 2): Filter IS NOT NULL predicates from scan on client

Repository: kudu Updated Branches: refs/heads/master 6d696bcf7 -> 1a60ed7d2 KUDU-1652 (part 2): Filter IS NOT NULL predicates from scan on client Part 1 fixed the C++ client/server and Java client so that specifying a an IS NOT NULL predicate on a non-nullable column would not result in a

kudu git commit: KUDU-1652 (part 2): Filter IS NOT NULL predicates from scan on client

Repository: kudu Updated Branches: refs/heads/branch-1.0.x 40637ef8f -> d2ebf5aa7 KUDU-1652 (part 2): Filter IS NOT NULL predicates from scan on client Part 1 fixed the C++ client/server and Java client so that specifying a an IS NOT NULL predicate on a non-nullable column would not result in

[2/3] kudu git commit: [docs] Cleanup beta mentions, links

[docs] Cleanup beta mentions, links Change-Id: I56fa65be9d027f5b40628aa5db5f9e4a57c1fbc9 Reviewed-on: http://gerrit.cloudera.org:8080/4565 Reviewed-by: David Ribeiro Alves Tested-by: Kudu Jenkins Reviewed-on: http://gerrit.cloudera.org:8080/4663 Reviewed-by: Dan Burkert

[3/3] kudu git commit: [docs] - Impala doc - partitioning required

[docs] - Impala doc - partitioning required It seems a blurb in the impala page got missed when partitioning became required on table create. Updated verbiage without removing reference links. Change-Id: I022b49c4e0b17031b9f14648ef0216ba0094b92b Reviewed-on: http://gerrit.cloudera.org:8080/4595

kudu git commit: Additional 1.0.1 release notes

Repository: kudu Updated Branches: refs/heads/branch-1.0.x 5cbcffa59 -> 6e1726219 Additional 1.0.1 release notes Change-Id: I5316a66d8d40e4ea532ba378868fbfe26864d52d Reviewed-on: http://gerrit.cloudera.org:8080/4665 Tested-by: Dan Burkert Reviewed-by: David Ribeiro Alves

kudu git commit: KUDU-1681: DNS resolution failure of master hostname causes tserver crash

Repository: kudu Updated Branches: refs/heads/branch-1.0.x deee2f184 -> 8143d1ff4 KUDU-1681: DNS resolution failure of master hostname causes tserver crash Also changes the fail_dns_resolution flag to make DNS resolution return an error Status instead of an empty resolved list. From the

[4/5] kudu git commit: Document Impala and Spark integration known issues & limitations

Document Impala and Spark integration known issues & limitations Change-Id: I993a09a00f5ab0049fec95e967abc1740b44dc8d Reviewed-on: http://gerrit.cloudera.org:8080/4443 Tested-by: Dan Burkert Reviewed-by: Jean-Daniel Cryans (cherry picked from commit

[5/5] kudu git commit: [doc] KUDU-1630 impala_tables.html link is showing as text and not actual link

[doc] KUDU-1630 impala_tables.html link is showing as text and not actual link Missing space before the hyperlink was causing this issue. Hyperlinked "Impala documentation" text to appropriate impala link. Complete text being: See the Impala documentation for more information about internal and

kudu git commit: [c++ client] notes for timestamp-related methods

Repository: kudu Updated Branches: refs/heads/branch-1.0.x 6e1726219 -> 120b1235e [c++ client] notes for timestamp-related methods Added notes on anticipated changes on methods related to getting and setting hybrid timestamps while performing READ_AT_SNAPSHOT scan operations: *

kudu git commit: [java client] Tight-ish loop in master lookups if a tablet doesn't have a leader

Repository: kudu Updated Branches: refs/heads/master 1a60ed7d2 -> d87486c47 [java client] Tight-ish loop in master lookups if a tablet doesn't have a leader There's currently the possibility of a situation like this: 1. sendRpcToTablet finds the tablet that the RPC is going to doesn't have a

kudu git commit: Apache Kudu 1.0.1 Release Notes

Repository: kudu Updated Branches: refs/heads/branch-1.0.x d2ebf5aa7 -> deee2f184 Apache Kudu 1.0.1 Release Notes Change-Id: I16986205e8c29dd42393b4e1b43d44a3fdb4daca Reviewed-on: http://gerrit.cloudera.org:8080/4621 Tested-by: Dan Burkert Reviewed-by: Todd Lipcon

[2/2] kudu git commit: cli tool: List all tablets/replica_uuids with 'kudu table list'

cli tool: List all tablets/replica_uuids with 'kudu table list' I noticed that given a tablet or replica_uuid we need to execute multiple nested commands and also need to correlate tablets and their replica uuids and their relation to tables. Added a verbose flag to 'kudu table list' to make this

kudu git commit: [java-client] address follow-up comments to c7dab4

Repository: kudu Updated Branches: refs/heads/master 1961305d8 -> dd49515f3 [java-client] address follow-up comments to c7dab4 Change-Id: Id9c071b8c267ed86f74c6d13b49d4c4693fddc94 Reviewed-on: http://gerrit.cloudera.org:8080/4442 Reviewed-by: Jean-Daniel Cryans

kudu git commit: make election timeout jitter more aggressive

Repository: kudu Updated Branches: refs/heads/master dd49515f3 -> b069dd28b make election timeout jitter more aggressive Random election timeout jitter is necessary in Raft in order to guarantee that an election can be won. If the jitter is smaller than RTT or the accuracy of clocks, then

kudu git commit: Replace incubator-kudu links with kudu in docs

Repository: kudu Updated Branches: refs/heads/master d911b304d -> c1aecf217 Replace incubator-kudu links with kudu in docs A number of docs were referring to the old incubator link: https://github.com/apache/incubator-kudu As opposed to the new, non-incubator link.

kudu git commit: KuduSession: do not advertise thread-safety

Repository: kudu Updated Branches: refs/heads/master c1aecf217 -> 2976b2a2c KuduSession: do not advertise thread-safety Do not advertise thread-safety of KuduSession methods for the Kudu C++ client, even if they are thread-safe de facto. This is to have one set of semantics for both C++ and

[2/2] kudu git commit: [java-client] IN-list predicate

[java-client] IN-list predicate Change-Id: Ia556196963153f9db64d67bc699f96cb920ecac6 Reviewed-on: http://gerrit.cloudera.org:8080/4530 Tested-by: Kudu Jenkins Reviewed-by: Jean-Daniel Cryans Project: http://git-wip-us.apache.org/repos/asf/kudu/repo Commit:

[1/2] kudu git commit: KUDU-1363: Add IN-list predicate type

Repository: kudu Updated Branches: refs/heads/master 17d1367e1 -> cb92799ba KUDU-1363: Add IN-list predicate type Adds support in the C++ client for providing a set of equalities for a given column. Support for using IN list predicates from the Java client will be in a follow-up commit.

[2/3] kudu git commit: Replace some TSAN suppressions with annotations

Replace some TSAN suppressions with annotations In the case that there is a known race, we can either use an annotation or a suppression to avoid it causing issues. It turns out that using suppressions is rather inefficient: TSAN still needs to take the stack trace and feed it to the external

[1/3] kudu git commit: [util] shorten names for key/mapped types

Repository: kudu Updated Branches: refs/heads/master 1dba44167 -> 26ef33602 [util] shorten names for key/mapped types There are not functional changes in this change. Change-Id: I9126f9b9a4a2e227ded842eee85140f81051fba6 Reviewed-on: http://gerrit.cloudera.org:8080/4584 Tested-by: Kudu

[3/3] kudu git commit: Clean up switch statements in column_predicate.cc

Clean up switch statements in column_predicate.cc Where possible, remove default branches in favor of LOG(FATAL) calls outside the switch. This makes it more likely that the switch will be updated in the future for new predicate types, since C++ compilers can warn on incomplete switches if there

[4/8] kudu git commit: [java client] make DateFormat safe to use

[java client] make DateFormat safe to use Todd's error-prone patch discovered this problem in RowResult where we use a static DateFormat from multiple threads. The recommended way to do this is to use thread locals. Change-Id: I6d18ba34db0a0782fd0b7401e9aea7ae59c6b6f6 Reviewed-on:

[8/8] kudu git commit: cache: fix behavior on single-CPU systems

cache: fix behavior on single-CPU systems On a system with only a single CPU, shard_bits_ would be set to 0. This would then result in calculating 'hash >> (32 - 0)' which is undefined behavior. With optimizations, this would turn into a no-op, and we'd end up using the whole hash as the shard

[6/8] kudu git commit: KUDU-1651: Check empty dict before decoder eval

KUDU-1651: Check empty dict before decoder eval A completely null dictionary-encoded column will yield an empty dictionary. When decoder-level evaluation is on, this will result in an error when creating the bitmap of matching codewords, as bitmaps must have a non-zero size. In this scenario, it

[1/8] kudu git commit: KUDU-1090: relax MemTracker uniqueness constraint

Repository: kudu Updated Branches: refs/heads/branch-1.0.x 6f6e49ca9 -> 40637ef8f KUDU-1090: relax MemTracker uniqueness constraint This patch relaxes the MemTracker uniqueness constraint to avoid certain rare crashes (e.g. web UI takes a reference to a MemRowSet tracker during

[3/8] kudu git commit: [java client] Fix an NPE in KuduException

[java client] Fix an NPE in KuduException Saw this in a Jenkins run and also running ITClient on my machine. Change-Id: Iceddc6931e8d3a8cb807657fc5c0804f7052e48f Reviewed-on: http://gerrit.cloudera.org:8080/4488 Reviewed-by: Adar Dembo Tested-by: Kudu Jenkins (cherry picked

[7/8] kudu git commit: KUDU-1652: Partition pruning fails with IS NOT NULL predicate on PK column

KUDU-1652: Partition pruning fails with IS NOT NULL predicate on PK column Scan optimization could also fail in certain cases, but the fix is the same. Change-Id: Icba2defe74e86b0e266e668974bce0ad9b0e Reviewed-on: http://gerrit.cloudera.org:8080/4541 Reviewed-by: Adar Dembo

kudu git commit: Retry on early TimedOut error in ExternalMiniCluster::WaitForCatalogManager

Repository: kudu Updated Branches: refs/heads/master 41e0b0af2 -> a6d7a5c6d Retry on early TimedOut error in ExternalMiniCluster::WaitForCatalogManager WaitForCatalogManager loops sending a GetTableLocations RPC to a master waiting for it to be initialized. Previously, we would early return

[1/4] kudu git commit: macOS: fix compile error introduced by 141de3377e6bdd

Repository: kudu Updated Branches: refs/heads/master 87c1a5f50 -> 279b72916 macOS: fix compile error introduced by 141de3377e6bdd Change-Id: Id2a9556810feeb3128b97d60afb18a2547e0af5c Reviewed-on: http://gerrit.cloudera.org:8080/5009 Reviewed-by: Todd Lipcon Tested-by: Kudu

[3/4] kudu git commit: Enforce that REINSERTs are not supported in DeltaMemStores

Enforce that REINSERTs are not supported in DeltaMemStores We never really supported REINSERTs in the delta memstore but had code that seemed like we eventually would. This adds a CHECK to make sure we never see REINSERTs in the DeltaMemStore and also simplifies some of the code accordingly.

[4/4] kudu git commit: Add IN LIST release notes

Add IN LIST release notes Change-Id: I61a6a1871098a3252e09556ab510ca9693d14d74 Reviewed-on: http://gerrit.cloudera.org:8080/5010 Reviewed-by: Jean-Daniel Cryans Tested-by: Dan Burkert Project: http://git-wip-us.apache.org/repos/asf/kudu/repo Commit:

[2/4] kudu git commit: [rpc] fixed misprint in RegisterConnection

[rpc] fixed misprint in RegisterConnection Fixed misprint in ReactorThread::RegisterConnection: do not put already shut down connection into the server connection list. Change-Id: Ifad34fa64caabebe265aea3c629657046b6818b5 Reviewed-on: http://gerrit.cloudera.org:8080/5003 Tested-by: Kudu Jenkins

[2/2] kudu git commit: installation.adoc: Add openssl and Kerberos dependencies

installation.adoc: Add openssl and Kerberos dependencies This adds openssl header and kerberos dependencies to the from-source installation instructions. Change-Id: I45b4e06480e8654a3021a09f38ec9df2c5dd Reviewed-on: http://gerrit.cloudera.org:8080/4981 Reviewed-by: Adar Dembo

[1/2] kudu git commit: Misc. fixes for Kerberos compatibility on OS X

Repository: kudu Updated Branches: refs/heads/master ae07d0dc4 -> 2adb3aeb1 Misc. fixes for Kerberos compatibility on OS X This commit makes a few changes in order to have better compatibility with the system macOS Heimdal kerberos: 1. krb5kdc now uses UDP instead of TCP; the heimdal client

[3/5] kudu git commit: Commit CSD metrics for 1.0.0 and 1.0.1

http://git-wip-us.apache.org/repos/asf/kudu/blob/0c52f161/java/kudu-csd/old-version-metrics/1.0.1-master.json -- diff --git a/java/kudu-csd/old-version-metrics/1.0.1-master.json

[1/5] kudu git commit: add 1.0.1 release notes to prior_release_notes.adoc

Repository: kudu Updated Branches: refs/heads/master d065e3316 -> 0c52f161b add 1.0.1 release notes to prior_release_notes.adoc Change-Id: Ieff57fd7120e98e2e1a7c0812e6c04eca704c6ad Reviewed-on: http://gerrit.cloudera.org:8080/4699 Tested-by: Kudu Jenkins Reviewed-by: Jean-Daniel Cryans

[2/5] kudu git commit: Commit CSD metrics for 1.0.0 and 1.0.1

http://git-wip-us.apache.org/repos/asf/kudu/blob/0c52f161/java/kudu-csd/old-version-metrics/1.0.1-tserver.json -- diff --git a/java/kudu-csd/old-version-metrics/1.0.1-tserver.json

[5/5] kudu git commit: Commit CSD metrics for 1.0.0 and 1.0.1

Commit CSD metrics for 1.0.0 and 1.0.1 Change-Id: If7e9867289522e0c0c04bcbffaf04fb249589287 Reviewed-on: http://gerrit.cloudera.org:8080/4695 Tested-by: Kudu Jenkins Reviewed-by: Todd Lipcon Project: http://git-wip-us.apache.org/repos/asf/kudu/repo Commit:

kudu git commit: MiniKdc for Java

Repository: kudu Updated Branches: refs/heads/master 4b5425aa3 -> 1a8ce4269 MiniKdc for Java Change-Id: Ie24eaa94fae14ca91fb4fdd2deae1f9aec58438b Reviewed-on: http://gerrit.cloudera.org:8080/4788 Tested-by: Kudu Jenkins Reviewed-by: Alexey Serbin Project:

[3/3] kudu git commit: Fix MiniKdc dtor crash

Fix MiniKdc dtor crash The C++ MiniKdc had a bug that caused a CHECK failure when the dtor ran while stopped. Change-Id: Iccfa5cdbfadd82d021e9ccede020e1d19e7054fc Reviewed-on: http://gerrit.cloudera.org:8080/4839 Reviewed-by: Adar Dembo Tested-by: Dan Burkert

[1/3] kudu git commit: Add a design doc for rpc retry/failover semantics

Repository: kudu Updated Branches: refs/heads/master ff24651cb -> 6ed6989ab Add a design doc for rpc retry/failover semantics This adds the final version of the rpc retry/failover doc and includes details of the final implementation. It also includes a guide and hints on how to implement

[2/3] kudu git commit: Add krb5 dependency to Java README

Add krb5 dependency to Java README I think it's a little premature to add it to installation.adoc, since it isn't required for building or operating Kudu yet (it's a test-only dependency). Change-Id: I5e4ffe0bee242ddec3a9effd4c3404348f1b889c Reviewed-on: http://gerrit.cloudera.org:8080/4840

[6/6] kudu git commit: [java-client]: support for Kerberized RPC

[java-client]: support for Kerberized RPC This commit adds initial support for connecting to a Kerberized cluster with the Java client. The application is required to have an active login context which contains a subject with Kerberos credentials when creating the KuduClient to connect to a

[4/6] kudu git commit: Change TabletHistoryGCITest back to manual flushes

Change TabletHistoryGCITest back to manual flushes Mike found a potential problem with client flushes when there are concurrent manual and automatic flushes going on (KUDU-1761). Until we address that this changes this test back to MANUAL_FLUSH with a single flush at the end of each batch. This

[2/6] kudu git commit: Allow to skip the scan part of full_stack-insert-scan-test

Allow to skip the scan part of full_stack-insert-scan-test This test is good to microbenchmark/run perf on writes but the scan part of the test introduces noise in this case. This patch adds a flag to allow to skip the scan part of the test. Change-Id: I93e4c0dd5637bb8992f5f2fd51d4f3811e468974

[5/6] kudu git commit: c++11'ize CompactionInput descedants with 'override'

c++11'ize CompactionInput descedants with 'override' Change-Id: I9dd25733f0f33ff824b854cb4f31b4a5d973a685 Reviewed-on: http://gerrit.cloudera.org:8080/5227 Tested-by: Kudu Jenkins Reviewed-by: Alexey Serbin Reviewed-by: Todd Lipcon Project:

[1/6] kudu git commit: Remove suppression for already-fixed KUDU-189

Repository: kudu Updated Branches: refs/heads/master 11fa20723 -> dad80bd50 Remove suppression for already-fixed KUDU-189 The code mentioned in this suppression was removed in July 2014 by 0c440b0253e97209c6481608e941e6532db3bd20 Change-Id: I8cd3b917a25cb6d0de7e04d26087cf249f5d00d0

kudu git commit: Add KuduTable.getFormattedRangePartitions method

Repository: kudu Updated Branches: refs/heads/master 22067edb4 -> 6be3f328a Add KuduTable.getFormattedRangePartitions method This adds an Impala-specific, unstable API to the KuduTable class for retrieving the list of the range partitions in a table. The partitions are formatted according to

[3/6] kudu git commit: [java] Add checkstyle coverage to Java build

http://git-wip-us.apache.org/repos/asf/kudu/blob/22067edb/java/kudu-client/src/main/java/org/apache/kudu/client/TabletClient.java -- diff --git a/java/kudu-client/src/main/java/org/apache/kudu/client/TabletClient.java

[4/6] kudu git commit: [java] Add checkstyle coverage to Java build

http://git-wip-us.apache.org/repos/asf/kudu/blob/22067edb/java/kudu-client/src/main/java/org/apache/kudu/client/KuduScanToken.java -- diff --git a/java/kudu-client/src/main/java/org/apache/kudu/client/KuduScanToken.java

[2/6] kudu git commit: [java] Add checkstyle coverage to Java build

http://git-wip-us.apache.org/repos/asf/kudu/blob/22067edb/java/kudu-client/src/test/java/org/apache/kudu/client/TestPartitionPruner.java -- diff --git

  1   2   3   4   5   6   7   >