Repository: kudu
Updated Branches:
refs/heads/master 5f34a7d09 -> c9b9315a0
KUDU-2113 Segfault because of consensus conflict and missing tablet server
Sometimes the master doesn't report all tablet servers. This can
happen, for example, when the master is starting. When this
happened and
Add metrics for tablet state and tablet copy
This patch adds a metric to each tablet that records its state,
e.g. RUNNING or TOMBSTONED. Metrics systems will need to
aggregate the state strings to give server-level statistics on
tablet state.
This patch also adds metrics for tablet copy
Repository: kudu
Updated Branches:
refs/heads/master 655cf1d5e -> 15e28176f
Update version to 1.6.0-SNAPSHOT
Change-Id: I961d9804b34ecdc09884273dbcc62d0de1809d8d
Reviewed-on: http://gerrit.cloudera.org:8080/7869
Reviewed-by: Adar Dembo
Tested-by: Kudu Jenkins
Project:
Repository: kudu
Updated Branches:
refs/heads/branch-1.5.x [created] 655cf1d5e
Repository: kudu
Updated Branches:
refs/heads/master ed76e0b2c -> 655cf1d5e
[release notes] add new kudu tool features and metrics improvements
Change-Id: Ic9514999a913c52fdfc9ed3cf6a85da50b190c68
Reviewed-on: http://gerrit.cloudera.org:8080/7865
Reviewed-by: Adar Dembo
Repository: kudu
Updated Branches:
refs/heads/master 975cb5cdb -> ed76e0b2c
docs: Add release note for tombstoned voting
Change-Id: I14d06ac34ee43cd5592a5fd77074224997812b4e
Reviewed-on: http://gerrit.cloudera.org:8080/7858
Tested-by: Mike Percy
Reviewed-by: Dan Burkert
Repository: kudu
Updated Branches:
refs/heads/master 334dcbe2b -> 975cb5cdb
docs: add 1.5.0 release note for tablet copying and
flushing improvements
Change-Id: Idf254bfcad27b66e8fa515be92285c876332a0be
Reviewed-on: http://gerrit.cloudera.org:8080/7862
Tested-by: Kudu Jenkins
Reviewed-by: Dan
Repository: kudu
Updated Branches:
refs/heads/master c7a22fa6b -> 31ddbcf0b
docs: add 1.5.0 release note for disk failures
Change-Id: I3ad45527dc32a4946c817ddbef2149433f3b5b89
Reviewed-on: http://gerrit.cloudera.org:8080/7856
Reviewed-by: Dan Burkert
Tested-by: Dan
Repository: kudu
Updated Branches:
refs/heads/master faa0b14ef -> 1c70e5df8
KUDU-1544: Race in Java client's AsyncKuduSession.apply()
This fixes a potential race in AsyncKuduSession.apply by acquiring the
notification deferred early. See the JIRA for details about the race.
Acquiring the
release notes: add contributor counts
Change-Id: Ib385b93671056c37474367266291370d7893550c
Reviewed-on: http://gerrit.cloudera.org:8080/7848
Reviewed-by: Todd Lipcon
Tested-by: Todd Lipcon
Project: http://git-wip-us.apache.org/repos/asf/kudu/repo
Commit:
release notes: add web UI config dashboard entry
Change-Id: I8812e1643afc50af0633343aa76fbea79bc83d72
Reviewed-on: http://gerrit.cloudera.org:8080/7847
Tested-by: Kudu Jenkins
Reviewed-by: Todd Lipcon
Project: http://git-wip-us.apache.org/repos/asf/kudu/repo
Commit:
Repository: kudu
Updated Branches:
refs/heads/master 2895538e9 -> 9a07b2fed
KUDU-2102: fix PosixRWFile::Sync to guarantee durability when
used concurrently
PosixRWFile has an AtomicBool to optimize the perfomance of Sync()
so that when two threads call Sync() at the same time, only one will
Repository: kudu
Updated Branches:
refs/heads/master 64c0b15e9 -> 2895538e9
[release-notes] KUDU-2013 is fixed in Kudu 1.5
Change-Id: Ie3098b05747bc19b5d1a8f931d171cd4048c8bbf
Reviewed-on: http://gerrit.cloudera.org:8080/7840
Reviewed-by: Todd Lipcon
Reviewed-by: Dan Burkert
KUDU-1726: Avoid fsync-per-block in tablet copy
This patch incorporates BlockTransaction API into tablet
copy, to avoid fsync() on each block during copying. Blocks
are synced to disk together once the tablet copy is complete.
It also introduces a new block manager metric 'total_disk_sync'
to
Repository: kudu
Updated Branches:
refs/heads/master 81fad4da9 -> 3d42a264c
KUDU-1943: Add BlockTransaction to Block Manager
This adds a new layer of abstraction 'BlockTransaction' to
Block Manager, to accumulate new blocks in a logical operation,
e.g. flush/compaction. Avoid using fsync()
Repository: kudu
Updated Branches:
refs/heads/master f9a6a1b2d -> 81fad4da9
gradle: target JRE 7 compatibility
Change-Id: Ib836a748f0f000b07460dd3141940655dbfb7a29
Reviewed-on: http://gerrit.cloudera.org:8080/7785
Tested-by: Kudu Jenkins
Reviewed-by: Todd Lipcon
Reviewed-by:
Repository: kudu
Updated Branches:
refs/heads/master 7986ee028 -> fa2757763
[java] Include Spark/Scala base version in kudu-spark-tools
Renames the kudu-spark-tools artifact to include the
spark and scala base version. Including the scala
version is standard practice and helps users avoid
Repository: kudu
Updated Branches:
refs/heads/master 055425245 -> 84096d89b
KUDU-2104. Upgrade to Spark 2.2.0
Upgrades to Spark 2.2.0 and documents the Java 8
runtime requirement.
Also adjusts the Jenkins scripts to use Java 8 when
building and running.
Change-Id:
Repository: kudu
Updated Branches:
refs/heads/master 7eb58663f -> ecd67486b
Web UI: switch /maintenance-manager endpoint to mustache
This cherry-picks just the changes to switch the /maintenance-manager
endpoint to use mustache by Sam Okrent in
https://gerrit.cloudera.org/#/c/7570/, with the
Repository: kudu
Updated Branches:
refs/heads/master cb9a6b84d -> 9285f2b44
Add security dashboard to web UI
Adds a new "Configuration" page to the web UI
that displays a list of security configuration
options, their current values, whether or not
they're the most secure settings, and
http://git-wip-us.apache.org/repos/asf/kudu/blob/cb9a6b84/www/jquery-3.2.1.min.js
--
diff --git a/www/jquery-3.2.1.min.js b/www/jquery-3.2.1.min.js
new file mode 100644
index 000..644d35e
--- /dev/null
+++
http://git-wip-us.apache.org/repos/asf/kudu/blob/cb9a6b84/www/bootstrap/css/bootstrap.css
--
diff --git a/www/bootstrap/css/bootstrap.css b/www/bootstrap/css/bootstrap.css
deleted file mode 100644
index 1b519e2..000
---
http://git-wip-us.apache.org/repos/asf/kudu/blob/cb9a6b84/www/bootstrap/img/glyphicons-halflings-white.png
--
diff --git a/www/bootstrap/img/glyphicons-halflings-white.png
b/www/bootstrap/img/glyphicons-halflings-white.png
Update web UI style
Update to latest versions of jquery and bootstrap,
and make some small stylistic changes to the web UI.
Change-Id: I8f9de6f5e555e33cbc2e25801cf59b74640d83ee
Reviewed-on: http://gerrit.cloudera.org:8080/7504
Tested-by: Kudu Jenkins
Reviewed-by: Dan Burkert
Repository: kudu
Updated Branches:
refs/heads/master 0fe3a288d -> cb9a6b84d
Integrate Mustache templates to webserver
This commit alters the design of webserver path handlers
to use Mustache templates instead of generating HTML
in server code. Path handlers now accept a json object
as a
http://git-wip-us.apache.org/repos/asf/kudu/blob/cb9a6b84/www/bootstrap/js/bootstrap.min.js
--
diff --git a/www/bootstrap/js/bootstrap.min.js
b/www/bootstrap/js/bootstrap.min.js
index 55da8c3..9bcd2fc 100644
---
http://git-wip-us.apache.org/repos/asf/kudu/blob/cb9a6b84/www/jquery-1.11.1.min.js
--
diff --git a/www/jquery-1.11.1.min.js b/www/jquery-1.11.1.min.js
deleted file mode 100644
index ab28a24..000
--- a/www/jquery-1.11.1.min.js
Repository: kudu
Updated Branches:
refs/heads/master 27854fd49 -> be8e3c22b
thirdparty: use ninja when possible
Using ninja to build LLVM, etc, is a bit faster than using make, and
produces less ugly output.
Change-Id: Idacd6d8bc9ad371c3c925f7c628ad69785bb870e
Reviewed-on:
Repository: kudu
Updated Branches:
refs/heads/master 69ee17dc7 -> 81939782a
Add json builder utility
The EasyJson class serves as a wrapper around rapidjson
Value objects, to simplify their usage. It is intended solely
for building json objects, not writing/parsing.
Simplifies code like
Repository: kudu
Updated Branches:
refs/heads/master bce46e12f -> 69ee17dc7
java: refactor ITClientStress for more reusable code
Change-Id: I263b5681b52df0fa496cce2f58c0605c66810e64
Reviewed-on: http://gerrit.cloudera.org:8080/7360
Tested-by: Kudu Jenkins
Reviewed-by: Jean-Daniel Cryans
util: add Status::AndThen combinator
The AndThen combinator makes it easier to chain together
Status-returning operations when failure recovery is something other
than returning early. RETURN_NOT_OK already handles the return-early
case elegantly.
Also tweaks Status::CloneAndPrepend and
Repository: kudu
Updated Branches:
refs/heads/master bbed78cbd -> 15e33c122
[java] update dependencies and clean up some deprecate method usage
Change-Id: I4b9746f70187c5a7ca0dd016db92ae4e6de7ecc3
Reviewed-on: http://gerrit.cloudera.org:8080/7276
Tested-by: Kudu Jenkins
Reviewed-by: Dan
Repository: kudu
Updated Branches:
refs/heads/master 5a258508f -> 57a07ae72
KUDU-1929: [rpc] Allow using encrypted private keys for TLS
* This patch adds a new flag for a "password command" for the RPC
private key.
* This also makes the webserver move to using the new function
Repository: kudu
Updated Branches:
refs/heads/master a85b294fb -> 5a258508f
[java] Ensure the shaded packages are always in the same location
Currently when a module depends on kudu-clients and is
shaded it is possible to get bloat in the jar due to
dependencies like guava being included in
Repository: kudu
Updated Branches:
refs/heads/master 52e63d8aa -> a85b294fb
[java] remove unused scalamock dependency
Change-Id: I97005bd5211abd0b119373ec265b8e9d01b385cf
Reviewed-on: http://gerrit.cloudera.org:8080/7292
Tested-by: Kudu Jenkins
Reviewed-by: Dan Burkert
Repository: kudu
Updated Branches:
refs/heads/master baa8d1716 -> ab07952fb
Add mustache as a dependency
Mustache is a popular template engine. The version
added by this commit is a C++ implementation found
at https://github.com/henryr/cpp-mustache. Mustache
is intended to be use for
Repository: kudu
Updated Branches:
refs/heads/master f9e51ca63 -> 5121df65e
KUDU-1955 refuse to use world-readable keytabs
Allowing users to supply keytab files and TLS private keys
with world-readable permissions lessens a cluster's security.
During Kerberos/TLS initialization, servers now
Repository: kudu
Updated Branches:
refs/heads/master 870343f0b -> a5b7ece1b
fs_manager: clean tmp files from all special directories
Commit c26e09e regressed tmp file cleanup in that the tablet-meta and cmeta
directories were no longer cleaned. This is because I assumed that the
cleanup done
maven: remove unused repositories, plugins, and dependencies
Change-Id: I06513b301923a89334cd8820b5cd4abf11543425
Reviewed-on: http://gerrit.cloudera.org:8080/7201
Reviewed-by: Adar Dembo
Tested-by: Kudu Jenkins
Project: http://git-wip-us.apache.org/repos/asf/kudu/repo
Fix all 'misc-string-compare' warnings from clang-tidy
See https://clang.llvm.org/extra/clang-tidy/checks/misc-string-compare.html
for details
(auto-generated using clang-tidy --fix with only this check enabled)
Change-Id: I18e609172eda7793f79ed9bb8ddafdf3b85d5fb0
Reviewed-on:
Repository: kudu
Updated Branches:
refs/heads/master 6dade2aa0 -> 8d2ca982f
Fix various use-after-move errors
This fixes all of the cases where clang-tidy detects
'misc-use-after-move'. This is necessary before enabling protobuf move
constructors to avoid test failures.
Change-Id:
Enable move constructors for protobufs
This enables an experimental option for protobuf to generate move
constructors for protos. This should make storing protobufs in
containers more performant, and also should allow us to avoid some
current patterns where we use ::Swap() instead of relying on
Tweak build-and-test.sh to fix spark1 test failures
Maven seems to be caching old artifacts which is leading to
hard-to-diagnose test failures in some branches. This commit changes the
spark1 test to do a full clean/rebuild instead of selectively removing
the kudu-spark target dir. I've also
Repository: kudu
Updated Branches:
refs/heads/master 0f31386d2 -> f329e089f
Bump googletest
The new clang tidy is issuing warnings about gtest macros, the new
version seems to fix the issue.
Change-Id: I1cf6eee7007857f66b4177d3bfed665d8702e37f
Reviewed-on:
Repository: kudu
Updated Branches:
refs/heads/branch-1.4.x afcd412ba -> 9adbdd771
[tools] Use PrintTable to format ksck's consensus matrix
The first version of ksck's consensus matrix used its own table
code, but actually there was already nice table code available
in tool_action_common.h.
Repository: kudu
Updated Branches:
refs/heads/branch-1.3.x a1234c05a -> 713703ea1
[java-client] update protoc maven plugin
The current protoc maven plugin is not hosted on maven central
and the hosting domain is unavailable. It also has not been
updated since 2012.
It appears like the most
Repository: kudu
Updated Branches:
refs/heads/master e3b7d4dc1 -> ce83b6cef
Bump LLVM to 4.0.0
I'm working on integrating Thrift into the Kudu build, and hit
https://reviews.llvm.org/D22800, which is fixed in 4.0.0.
Change-Id: I64f98bb56d2510ba03ef5ea477a4d7bb4b48c555
Reviewed-on:
Repository: kudu
Updated Branches:
refs/heads/master 7e02ede33 -> 86fac3fa2
KUDU-1911 improve missing required arg message
When using a tool that has a required positional argument, excluding that
argument or passing it as a flag caused a confusing error message to be
printed out ("must
Repository: kudu
Updated Branches:
refs/heads/master e2097c8b9 -> 12683435f
KUDU-1988: add support for advertised host:port info.
Prior to this change, it was difficult to deploy a Kudu cluster in
containerized environments where each container has a unique, local
subnetwork with an
Repository: kudu
Updated Branches:
refs/heads/master d84fc1072 -> c5fe583e9
Fix flaky test TestRecoverFromOpIdOverflow (again)
The previous attempt to fix this in commit
f0580499dc50e8a47ff6251301cdc15b9b79edcb had a flaw, but this test
really does fix the primary source of the flakiness.
Repository: kudu
Updated Branches:
refs/heads/master fb6cc895b -> d84fc1072
KUDU-2020: tserver failure causes multiple tablet copy operations per
under-replicated tablet
The 'active ingredient' in this patch is the change to
TsTabletManager::StartTabletCopy that causes an ALREADY_INPROGRESS
Repository: kudu
Updated Branches:
refs/heads/master 95b52 -> d354693c1
docs: revise security doc based on KUDU-1875 and
add release notes for 1.4
Change-Id: I2e386203f5ed3ef66e2ec136e67738b8c7eb8b1a
Reviewed-on: http://gerrit.cloudera.org:8080/6922
Reviewed-by: Dan Burkert
KUDU-2017. Fix calculation of perf improvement for flushes
The code to calculate the MM "perf improvement score" for a flush had two bugs:
(1) we calculated threshold - current_usage instead of current_usage -
threshold, which resulted in a negative score.
(2) we had an unsigned integer
KUDU-1998. Re-enable tcmalloc on OSX
We switched back away from using malloc hooks for memory tracking, so
this shouldn't be an issue anymore.
Change-Id: I9bb3329aa4c12595dedf27d022558671e90206fb
Reviewed-on: http://gerrit.cloudera.org:8080/6917
Reviewed-by: Adar Dembo
Repository: kudu
Updated Branches:
refs/heads/master c178563f6 -> 4347174d7
process_memory: go back to non-incremental tracking
The incremental tracking has more overhead than expected, even with our
fancy striped counters, etc. In tpch_real_world benchmarks, the
LongAdder::IncrementBy() and
[c++ client] less logging about authn token
LOG(INFO) --> VLOG(2) on receiving and adopting authn token by
the Kudu C++ client.
Change-Id: I13693bc89e528eb4dfdd0e85d6d5b8faac8edadf
Reviewed-on: http://gerrit.cloudera.org:8080/6914
Tested-by: Kudu Jenkins
Reviewed-by: Todd Lipcon
Fix broken url to "Building a Better Bloom Filter"
Change-Id: Ic113677c7ff4c2080c0ac07cb8f90db77aaeaa52
Reviewed-on: http://gerrit.cloudera.org:8080/6920
Reviewed-by: Dan Burkert
Tested-by: Dan Burkert
Project:
Repository: kudu
Updated Branches:
refs/heads/master 4af33ba13 -> b28a59266
spark: add support for fault tolerant scanner
This adds support to use fault tolerant scanner for spark job.
By default non fault tolerant scanner is used. To turn on fault
tolerant scanner, use job config:
[java client] removes duplicate checks at KuduScanToken
KUDU-579 added fault tolerant scanner support at java
client. However,it missed TODO for KUDU-1040 and results
in duplicate code at KuduScanToken. This patch removes it.
Change-Id: I7e036d5fe333009a74294d27c54fe846cdc404e9
Reviewed-on:
Repository: kudu
Updated Branches:
refs/heads/master a7e8994e4 -> 4ec6280c9
Bump glog to 0.3.5
Change-Id: I9f638970abfd1aa214cf0032b46064ff7edf4dc4
Reviewed-on: http://gerrit.cloudera.org:8080/6871
Tested-by: Kudu Jenkins
Reviewed-by: Dan Burkert
Project:
Repository: kudu
Updated Branches:
refs/heads/master b2d247fa4 -> a7e8994e4
cmake: fix protoc dependency typo
The effect of this typo was that protoc would always be statically
linked. Since it's only a build-time dependency, we never noticed
previously.
Change-Id:
Repository: kudu
Updated Branches:
refs/heads/branch-1.3.x 1aa34a1d2 -> ea87fdd8e
KUDU-1999: Spark connector should login with Kerberos credentials on driver
Tested on a secure cluster using the Spark ITBLL job:
spark2-submit \
--deploy-mode=cluster \
--master=yarn \
Repository: kudu
Updated Branches:
refs/heads/master c4c796a25 -> a877566e9
KUDU-1999: Spark connector should login with Kerberos credentials on driver
Tested on a secure cluster using the Spark ITBLL job:
spark2-submit \
--deploy-mode=cluster \
--master=yarn \
--principal=dan \
[c++-client] clear non-covered entries from meta cache on table open
Clearing non-covered range entries from the meta cache on table open
makes it easier to share client instances among different contexts
without having to worry about polluting the non-covering range cache.
Change-Id:
Repository: kudu
Updated Branches:
refs/heads/master 7b582a896 -> fdc022fe6
[fs] fix build on OS X
Added kudu_fs dependency for kudu_fs_test_util library.
Prior to this patch, the following error was reported if trying to build
on OS X:
Undefined symbols for architecture x86_64:
KUDU-579 [java_client] Scanner fault tolerance
This patch adds java client support to restart scanners if they
fail in the middle of table scanning. AsyncKuduScanner records
the final primary key retrieved in the previous batch. If a tserver
fails while scanning, the client marks the tserver as
Add 'kudu master list' tool
This also sneaks in a couple other changes:
- kudu tserver list now has a 'heartbeat' column option to display the
number of ms since the last heartbeat.
- A fix for the pretty table formatter.
Change-Id: I9e77d41b244143d1258b26b50c4d2a68dedd8f00
Reviewed-on:
Repository: kudu
Updated Branches:
refs/heads/master b2843ee5e -> 319a1bc15
Add 'kudu tserver list' tool
This adds a new tool action to list tablet servers, and associated
information. There are a few output formatting options: the default is a
human-readable table in the psql style; the
Repository: kudu
Updated Branches:
refs/heads/branch-1.3.x 3211d9781 -> 73918d055
[docs] Revise disk-failure-recovery workflow
Missed a few comments in the original commit.
Change-Id: I35c2ca56fdcf07917e7e0ba7f0d9be7b6bb19568
Reviewed-on: http://gerrit.cloudera.org:8080/6608
Reviewed-by:
htly_etl_service_account
+--admin-acl=hadoopadmin
+```
+
+Further information about these flags can be found in the configuration
+flag reference.
+// TODO(todd) add a link
+
+
+[[known-limitations]]
+== Known Limitations
+
+Kudu has a few known security limitations:
+
+// TODO(danburkert): add JIR
Repository: kudu
Updated Branches:
refs/heads/master bd24f04fb -> 9d03677e4
Add ksck section to admin guide common workflows
I've often wanted this when helping people through ksck.
Change-Id: I9631337b113d2c67be0057f728c68f792e8a4fd6
Reviewed-on: http://gerrit.cloudera.org:8080/6598
KUDU-1966: Data directories can be removed erroneously
Also revises the error messages in
PathInstanceMetadataFile::CheckIntegrity to be in terms of data
directories, since this is what end-users will be familiar with. These
errors should be rare, but they can happen if a user is monkeying around
Repository: kudu
Updated Branches:
refs/heads/master b95d3a4fa -> f304dd32f
Add --redact=none flag option for specifying no redaction
It's still possible to specify no redaction via --redact='', but I was
only able to determine how to do that by reading the code. I think this
makes it a bit
Repository: kudu
Updated Branches:
refs/heads/master 27b3de7ab -> eca517108
generic_iterators: avoid holding completed iters in merge/union
Previously, the MergeIterator and UnionIterator would hold on to all of
the iterators they consume until the entire process completed. Each such
iterator
Repository: kudu
Updated Branches:
refs/heads/master f4272eb5c -> 34310a8e9
rpc-bench: add ability to enable encryption
This adds a --enable-encryption flag for the benchmark which disables
the "negotiation-only" TLS optimization and enables SSL on the server.
This allows us to benchmark
Spark ITBLL
This adds a new end-to-end test for Spark based on the existing
MapReduce ITBLL test. Common parts of the two have been abstracted into
a utility class.
The implementation is only compatible with Spark2/Scala2.11, so the new
submodule is deactivated when compiling with the
kudu-spark: switch default profile from spark1/2.10 to spark2/2.11
Now that we have kudu-spark-tools which only works with Spark2, it's
probably best if developers are defaulted to running tests with Spark2.
This shouldn't affect CI or published artifacts.
Change-Id:
Repository: kudu
Updated Branches:
refs/heads/master bee8e875e -> 7206041f9
docs: Fix axis lable for hash-range-partitioning
Horizontal axis label should show range partition by time according to the
example not (host, metric)
Change-Id: I619d2082941de257cee4270079b46116944c77a6
Repository: kudu
Updated Branches:
refs/heads/master 80f951236 -> 4a33783ee
[consensus] add lock coverage for peer queue
A Peer or PeerQueue would lock without taking into account the
presence of a peer calling back after TabletCopy and updating its
members, which doesn't abide by external
Repository: kudu
Updated Branches:
refs/heads/branch-1.3.x dfd9b491f -> d784f04d8
[util] fixed env_util-test.cc compilation on OS X
The prior version failed to compile on OS X with clang version 3.9.0
(branches/release_39 277311) with error messaged containing:
Repository: kudu
Updated Branches:
refs/heads/master 5759a08d3 -> c5ec0ddb0
[macOS] fix security test failures
Change-Id: I5c4d464edc09f41151f95ad2f0e694f2bc00b0f0
Reviewed-on: http://gerrit.cloudera.org:8080/6346
Tested-by: Kudu Jenkins
Reviewed-by: Alexey Serbin
java: avoid spewing ClosedChannelException on client-initiated disconnects
When submitting an MR job, the submitting program opens a client
instance for a short amount of time, just to fetch an authentication
token. When it then calls 'close()', any TabletClient connections that
the client may
Repository: kudu
Updated Branches:
refs/heads/branch-1.3.x 86055e295 -> 27b850469
[rpc] Do not send error reply for invalid connection context
We recently introduced a nonce into the KRPC protocol when using
Kerberos for authentication. The client sends the encoded nonce to the
server in the
Fix Webserver option 'password_file' to pass correct Squeasel option
'global_auth_file' for enabling HTTP authorization.
Also add test case for web UI .htpasswd support.
Change-Id: I2d30f450abfb3d0addc0eef39bcf78c87e4298c5
Reviewed-on: http://gerrit.cloudera.org:8080/6300
Tested-by: Kudu Jenkins
[sys_catalog] no copying into intermediate KuduPartialRow
Use KuduPartialRow::SetStringNoCopy() instead of
KuduPartialRow::SetString() while setting fields for the system catalog
table. Copying the data into the intermediate KuduPartialRow object is
not necessary: the intermediate object is not
Repository: kudu
Updated Branches:
refs/heads/master 72895966c -> 937064f91
KUDU-1887: Add RpcContext::DiscardTransfer()
DiscardTransfer() allows RPC handlers to release the memory backing an
inbound call, without being forced to respond to the RPC at the same
time.
Change-Id:
Repository: kudu
Updated Branches:
refs/heads/master 97831eadf -> f24307db4
[exactly_once_rpc-test] Reduce flakyness of
TestExactlyOnceSemanticsGarbageCollection
This test is failing frequently in the last call with either:
Bad status: Network error: Recv() got EOF from remote (error 108)
or
[rpc] Do not send error reply for invalid connection context
We recently introduced a nonce into the KRPC protocol when using
Kerberos for authentication. The client sends the encoded nonce to the
server in the connection context message, to which the client does not
expect a reply. This commit
Repository: kudu
Updated Branches:
refs/heads/master f65feff68 -> 45548c90c
KUDU-1897: disable Kerberos replay cache
Change-Id: Ifbce55a0b12682fdf69e7b2c361c6336495db64d
Reviewed-on: http://gerrit.cloudera.org:8080/6254
Reviewed-by: Todd Lipcon
Tested-by: Kudu Jenkins
Repository: kudu
Updated Branches:
refs/heads/master abbd09a57 -> 55c3db350
Java Client: authenticate with login username when using SASL PLAIN
Also tweaks up some logging output I noticed when reviewing the
negotiation tests for this change.
Change-Id:
Repository: kudu
Updated Branches:
refs/heads/master 1716d61ce -> abbd09a57
Destabilize rpc_authentication and rpc_encryption flags
These flags are not yet ready to be stabilized, it's possible they will
go through some changes in the future.
Change-Id:
Repository: kudu
Updated Branches:
refs/heads/master 93b54f0ff -> 1716d61ce
Bump gflags to 2.2.0
Flags can now be specified with dashes instead of underscoresâ¼ï¸ï¸
Change-Id: I400dc8eb74cce47848497342dbe5f945e1a98461
Reviewed-on: http://gerrit.cloudera.org:8080/6207
Reviewed-by: Todd
Repository: kudu
Updated Branches:
refs/heads/branch-1.2.x 32d4d5f3c -> 1af814b09
KUDU-1851 - [python] TableAlterer direct instantiation causes SIGSEGV
The python client is currently crashing when a TableAlterer is directly
instantiated as opposed to instantiating via the new_table_alterer
Repository: kudu
Updated Branches:
refs/heads/master 0efc1e26a -> 0609277c4
KUDU-1901. Fix crash in concurrent OpenSSL usage
This fixes a crash in which a TLSContext's SSL_CTX was being used by a
thread to accept a connection at the same time as its underlying
certificate was being changed.
KUDU-1852. KuduTableAlterer should not crash with nullptr arguments
Change-Id: Ifa7c34b476111ecd33d4fd3ef5cc363a410ad76a
Reviewed-on: http://gerrit.cloudera.org:8080/5797
Reviewed-by: Dan Burkert
Tested-by: Kudu Jenkins
(cherry picked from commit
Workaround a leak in OpenSSL 1.0.0
This fixes another leak seen occasionally as a flaky test on RHEL 6:
Direct leak of 56 byte(s) in 1 object(s) allocated from:
#2 0x7f21e5fe037c in CRYPTO_malloc src/openssl/crypto/mem.c:306
#3 0x7f21e607f45c in EVP_PKEY_new
Repository: kudu
Updated Branches:
refs/heads/branch-1.2.x 7fa56a53a -> 32d4d5f3c
KUDU-1899. Fix support for empty string keys
We've never prevented users from inserting a row with "" as a key. However, a
faulty assertion in diskrowset.cc caused the tablet server to crash on flush in
this
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:
[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.
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,
301 - 400 of 606 matches
Mail list logo