[kudu-CR](gh-pages) Add blog post about multi-master support in Kudu 1.0
Jean-Daniel Cryans has posted comments on this change. Change subject: Add blog post about multi-master support in Kudu 1.0 .. Patch Set 1: Code-Review+2 Good stuff! -- To view, visit http://gerrit.cloudera.org:8080/3476 To unsubscribe, visit http://gerrit.cloudera.org:8080/settings Gerrit-MessageType: comment Gerrit-Change-Id: I40a513f6a616ed0d947837fd9f33f64f755465ff Gerrit-PatchSet: 1 Gerrit-Project: kudu Gerrit-Branch: gh-pages Gerrit-Owner: Adar Dembo Gerrit-Reviewer: Jean-Daniel Cryans Gerrit-Reviewer: Todd Lipcon Gerrit-HasComments: No
[kudu-CR] Add required Debian version to installation page
Jean-Daniel Cryans has posted comments on this change. Change subject: Add required Debian version to installation page .. Patch Set 1: > JD, are you ok with pushing this now with your recent deb 8 > testing? Yes. -- To view, visit http://gerrit.cloudera.org:8080/3389 To unsubscribe, visit http://gerrit.cloudera.org:8080/settings Gerrit-MessageType: comment Gerrit-Change-Id: I65cef323648d7934cb140046cf391d56867eeb56 Gerrit-PatchSet: 1 Gerrit-Project: kudu Gerrit-Branch: master Gerrit-Owner: Dan Burkert Gerrit-Reviewer: Adar Dembo Gerrit-Reviewer: Dan Burkert Gerrit-Reviewer: Jean-Daniel Cryans Gerrit-Reviewer: Kudu Jenkins Gerrit-HasComments: No
[kudu-CR] KUDU-1309: [java client] support tables with non-covering partition-key ranges
Dan Burkert has posted comments on this change. Change subject: KUDU-1309: [java client] support tables with non-covering partition-key ranges .. Patch Set 2: (32 comments) http://gerrit.cloudera.org:8080/#/c/3388/2/java/kudu-client/src/main/java/org/kududb/client/AsyncKuduClient.java File java/kudu-client/src/main/java/org/kududb/client/AsyncKuduClient.java: PS2, Line 139: This map :* is always the first to be updated, because that's the map from :* which all the lookups are done in the fast-path of the requests :* that need to locate a tablet. The second map to be updated is :* tablet2client, because it comes second in the fast-path :* of every requests that need to locate a tablet. The third map :* is only used to handle TabletServer disconnections gracefully. > Nit: reading this again (I reviewed it eons ago), I think this part of the I moved the non covered range cache out from under this comment since it really doesn't fit (it's not the same data indexed differently). I agree that this comment is not great, but it should be fixed with a proper meta cache. Line 155:* Currently once a non-covered range is added to the cache, it is never > Nit: prefix this part of the comment with "TODO:" since it is expected to c Done Line 576: Deferred openScanner(final AsyncKuduScanner scanner) { > True, this can be removed. Done Line 685: // Otherwise fall through to below where a GetTableLocations lookup will occur > Nit: end with a period. Done Line 1148: Pair nonCoveredRange = getNonCoveredRange(table.getTableId(), key); > Nit: if you want, you can call getTableId() once outside of the loop, store Done PS2, Line 1355: NonCoveredRangeCache nonCoveredRanges = nonCoveredRangeCaches.get(tableId); : if (nonCoveredRanges == null) { : nonCoveredRanges = new NonCoveredRangeCache(); : NonCoveredRangeCache oldCache = nonCoveredRangeCaches.putIfAbsent(tableId, nonCoveredRanges); : if (oldCache != null) { : nonCoveredRanges = oldCache; : } : } > How about just: putIfAbsent returns null if the entry was absent. Line 1378: nonCoveredRanges.addNonCoveredRange(EMPTY_ARRAY, firstStartKey); > Don't we need all of these calls to addNonCoveredRange() to be atomic? With It may contain stale entries, but that needs to be handled once add/drop partition is introduced anyway. http://gerrit.cloudera.org:8080/#/c/3388/2/java/kudu-client/src/main/java/org/kududb/client/AsyncKuduScanner.java File java/kudu-client/src/main/java/org/kududb/client/AsyncKuduScanner.java: Line 474: invalidate(); > Nit: invalidate() is called in both branches, can we pull it out? Done http://gerrit.cloudera.org:8080/#/c/3388/2/java/kudu-client/src/main/java/org/kududb/client/AsyncKuduSession.java File java/kudu-client/src/main/java/org/kududb/client/AsyncKuduSession.java: Line 127: private final List failedLookups = new ArrayList<>(); > I think you can use a CopyOnWriteArrayList here and omit the use of a lock I think in this case (where we are incrementally building up the list), a mutex is the best answer. Doing the checked put shuffle with an AtomicReference is going to be hard and involve spinning + rebuilding the list per spin. COWAL isn't even lock free: http://grepcode.com/file/repository.grepcode.com/java/root/jdk/openjdk/8u40-b25/java/util/concurrent/CopyOnWriteArrayList.java#433 Line 303: Callback>, ArrayList> { > Nit: not related to your change, but why is ArrayList the second type and n Done Line 306: batchResponses = batchResponses == null ? new ArrayList() : batchResponses; > Is a null batchResponses actually possible? How? I'm not sure, the code already handled the null case. When I comment out this line the tests still pass. JD, do you know what the deal is? PS2, Line 308: // flushTablet() can return null when a tablet we wanted to flush was already flushed. Those : // nulls along with BatchResponses are then put in a list by Deferred.group(). We first need : // to filter the nulls out. > OMG. Agreed. Refactored. Line 320: List responsesList = new ArrayList<>(size); > If we're going to go through the trouble of precomputing the size, we ought I thought about this, but the only way to figure out how many failed lookups there are is to take out the lock. Since there really shouldn't be failed lookups I decided not to slow down the normal path. I'm not too concerned about making the case where there are slow lookups be slow. Line 397: if (tablet == null) { > Why don't we perform this check to AUTO_FLUSH_SYNC on L377-384? Seems usefu The first step of sendRpcToTablet (again, d
[kudu-CR] KUDU-1309: [java client] support tables with non-covering partition-key ranges
Kudu Jenkins has posted comments on this change. Change subject: KUDU-1309: [java client] support tables with non-covering partition-key ranges .. Patch Set 3: Build Started http://104.196.14.100/job/kudu-gerrit/1975/ -- To view, visit http://gerrit.cloudera.org:8080/3388 To unsubscribe, visit http://gerrit.cloudera.org:8080/settings Gerrit-MessageType: comment Gerrit-Change-Id: Id65a1f8a95bb16fc0360a17021a391afd3c9d03f Gerrit-PatchSet: 3 Gerrit-Project: kudu Gerrit-Branch: master Gerrit-Owner: Dan Burkert Gerrit-Reviewer: Adar Dembo Gerrit-Reviewer: Dan Burkert Gerrit-Reviewer: Jean-Daniel Cryans Gerrit-Reviewer: Kudu Jenkins Gerrit-HasComments: No
[kudu-CR] KUDU-1309: [java client] support tables with non-covering partition-key ranges
Hello Jean-Daniel Cryans, Adar Dembo, Kudu Jenkins, I'd like you to reexamine a change. Please visit http://gerrit.cloudera.org:8080/3388 to look at the new patch set (#3). Change subject: KUDU-1309: [java client] support tables with non-covering partition-key ranges .. KUDU-1309: [java client] support tables with non-covering partition-key ranges This commit adds support for accessing and creating tables with non-covering range partitions to the Java client. The only public interface introduced is the CreateTableOptions.addRangeBound method. Otherwise, all the changes are necessary for reading or writing to tables with range partition gaps. Right now if a client attempts to write to a non-covered range, a NotFound status is returned, which matches the C++ client. JD pointed out that this makes it impossible for applications to distinguish between failed updates because the row doesn't exist, and failed updates because the partition range doesn't exist. I plan to fix this by introducing a new status code and fixing both clients in a follow up commit. Change-Id: Id65a1f8a95bb16fc0360a17021a391afd3c9d03f --- M java/kudu-client/src/main/java/org/kududb/client/AsyncKuduClient.java M java/kudu-client/src/main/java/org/kududb/client/AsyncKuduScanner.java M java/kudu-client/src/main/java/org/kududb/client/AsyncKuduSession.java M java/kudu-client/src/main/java/org/kududb/client/Batch.java M java/kudu-client/src/main/java/org/kududb/client/BatchResponse.java M java/kudu-client/src/main/java/org/kududb/client/Bytes.java M java/kudu-client/src/main/java/org/kududb/client/CreateTableOptions.java M java/kudu-client/src/main/java/org/kududb/client/CreateTableRequest.java M java/kudu-client/src/main/java/org/kududb/client/KuduRpc.java M java/kudu-client/src/main/java/org/kududb/client/KuduRpcResponse.java A java/kudu-client/src/main/java/org/kududb/client/NonCoveredRangeCache.java A java/kudu-client/src/main/java/org/kududb/client/NonCoveredRangeException.java M java/kudu-client/src/main/java/org/kududb/client/Operation.java M java/kudu-client/src/main/java/org/kududb/client/OperationResponse.java M java/kudu-client/src/main/java/org/kududb/client/Partition.java M java/kudu-client/src/main/java/org/kududb/client/RowError.java M java/kudu-client/src/main/java/org/kududb/client/RowResultIterator.java M java/kudu-client/src/test/java/org/kududb/client/BaseKuduTest.java M java/kudu-client/src/test/java/org/kududb/client/TestFlexiblePartitioning.java M java/kudu-client/src/test/java/org/kududb/client/TestKuduClient.java M java/kudu-client/src/test/java/org/kududb/client/TestKuduSession.java M java/kudu-client/src/test/java/org/kududb/client/TestKuduTable.java 22 files changed, 830 insertions(+), 138 deletions(-) git pull ssh://gerrit.cloudera.org:29418/kudu refs/changes/88/3388/3 -- To view, visit http://gerrit.cloudera.org:8080/3388 To unsubscribe, visit http://gerrit.cloudera.org:8080/settings Gerrit-MessageType: newpatchset Gerrit-Change-Id: Id65a1f8a95bb16fc0360a17021a391afd3c9d03f Gerrit-PatchSet: 3 Gerrit-Project: kudu Gerrit-Branch: master Gerrit-Owner: Dan Burkert Gerrit-Reviewer: Adar Dembo Gerrit-Reviewer: Dan Burkert Gerrit-Reviewer: Jean-Daniel Cryans Gerrit-Reviewer: Kudu Jenkins
[kudu-CR] [java-client] refactor AsyncKuduSession
Hello Jean-Daniel Cryans, Adar Dembo, I'd like you to do a code review. Please visit http://gerrit.cloudera.org:8080/3477 to review the following change. Change subject: [java-client] refactor AsyncKuduSession .. [java-client] refactor AsyncKuduSession This commit refactors how AsyncKuduSession handles buffering internally. Instead of using a buffer per tablet and a separate list of operations in tablet location lookup, we buffer all operations into a single large buffer, and then dispatch into per-tablet batches during flush. This result is simplified logic in AsyncKuduSession. AsyncKuduSession#apply is no longer reentrant, and lookups no longer need to be tracked explicitly. This patch was tested against YCSB to make sure that it does not result in performance regressions. With an increased buffer size of 5000 to match the 0.9 client's buffer size, the performance is largely the same. With this refactor the effective batch size is much smaller since mutationBufferSize now refers to the one buffer for all tablets instead of to a per-tablet buffer. Change-Id: I2ee6d029b1a56e254bfb9a870917883abeadb6b8 --- M java/kudu-client/src/main/java/org/kududb/client/AsyncKuduClient.java M java/kudu-client/src/main/java/org/kududb/client/AsyncKuduSession.java M java/kudu-client/src/main/java/org/kududb/client/Batch.java M java/kudu-client/src/main/java/org/kududb/client/Bytes.java M java/kudu-client/src/main/java/org/kududb/client/Operation.java M java/kudu-client/src/main/java/org/kududb/client/RowError.java M java/kudu-client/src/main/java/org/kududb/util/AsyncUtil.java M java/kudu-client/src/test/java/org/kududb/client/TestOperation.java 8 files changed, 587 insertions(+), 638 deletions(-) git pull ssh://gerrit.cloudera.org:29418/kudu refs/changes/77/3477/1 -- To view, visit http://gerrit.cloudera.org:8080/3477 To unsubscribe, visit http://gerrit.cloudera.org:8080/settings Gerrit-MessageType: newchange Gerrit-Change-Id: I2ee6d029b1a56e254bfb9a870917883abeadb6b8 Gerrit-PatchSet: 1 Gerrit-Project: kudu Gerrit-Branch: master Gerrit-Owner: Dan Burkert Gerrit-Reviewer: Adar Dembo Gerrit-Reviewer: Jean-Daniel Cryans
[kudu-CR] [java-client] refactor AsyncKuduSession
Kudu Jenkins has posted comments on this change. Change subject: [java-client] refactor AsyncKuduSession .. Patch Set 1: Build Started http://104.196.14.100/job/kudu-gerrit/1976/ -- To view, visit http://gerrit.cloudera.org:8080/3477 To unsubscribe, visit http://gerrit.cloudera.org:8080/settings Gerrit-MessageType: comment Gerrit-Change-Id: I2ee6d029b1a56e254bfb9a870917883abeadb6b8 Gerrit-PatchSet: 1 Gerrit-Project: kudu Gerrit-Branch: master Gerrit-Owner: Dan Burkert Gerrit-Reviewer: Adar Dembo Gerrit-Reviewer: Jean-Daniel Cryans Gerrit-Reviewer: Kudu Jenkins Gerrit-HasComments: No
[kudu-CR](gh-pages) Add blog post about multi-master support in Kudu 1.0
Hello Jean-Daniel Cryans, Todd Lipcon, I'd like you to do a code review. Please visit http://gerrit.cloudera.org:8080/3476 to review the following change. Change subject: Add blog post about multi-master support in Kudu 1.0 .. Add blog post about multi-master support in Kudu 1.0 I also fixed the placement of the tag in this week's update. Change-Id: I40a513f6a616ed0d947837fd9f33f64f755465ff --- M _posts/2016-06-21-weekly-update.md A _posts/2016-06-24-multi-master-1-0-0.md 2 files changed, 107 insertions(+), 2 deletions(-) git pull ssh://gerrit.cloudera.org:29418/kudu refs/changes/76/3476/1 -- To view, visit http://gerrit.cloudera.org:8080/3476 To unsubscribe, visit http://gerrit.cloudera.org:8080/settings Gerrit-MessageType: newchange Gerrit-Change-Id: I40a513f6a616ed0d947837fd9f33f64f755465ff Gerrit-PatchSet: 1 Gerrit-Project: kudu Gerrit-Branch: gh-pages Gerrit-Owner: Adar Dembo Gerrit-Reviewer: Jean-Daniel Cryans Gerrit-Reviewer: Todd Lipcon
[kudu-CR] Allow for reserving disk space for non-Kudu processes
Mike Percy has submitted this change and it was merged. Change subject: Allow for reserving disk space for non-Kudu processes .. Allow for reserving disk space for non-Kudu processes Adds gflags to reserve disk space such that Kudu will not use more than specified. Hadoop calls this functionality "du.reserved". If a WAL preallocation is attempted while the log disk is past its reservation limit, the log write will fail. As a result, RaftConsensus will cause the process to crash. The log block manager will use non-full disks if possible until all of the disks are full. If a flush or compaction is attempted when all disks are beyond their configured capacity then the LogBlockManager will return an error. As a result, the maintenance manager task will cause the process to crash. This initial implementation provides a "best effort" approach. Disk space checks are only done at preallocation time, and if writes continue beyond the preallocated point (for both a WAL segment and a data block) those writes will not be prevented. This makes it easier to provide a "friendly" option where the block manager will divert new writes to non-full disks, avoiding a hard crash when only one disk is past its reservation limit. In the future, we may want to add "hard" and "soft" limits, such that going beyond the soft limit will do what we do today, and going beyond the hard limit (say, by writing a very large data block past its preallocation point) will result in a crash. This patch includes: * Unit tests. * End-to-end test for flushing / compaction falling back to non-full disks due to disk space backpressure and finally crashing when there is no space left in any data dir. * End-to-end test for writes failing due to WAL disk space backpressure, causing a crash. Change-Id: Ifd0451d4dbddc1783019a53302de0263080939c7 Reviewed-on: http://gerrit.cloudera.org:8080/3135 Reviewed-by: Adar Dembo Tested-by: Mike Percy Reviewed-by: Jean-Daniel Cryans Reviewed-by: Todd Lipcon --- M src/kudu/consensus/log-test.cc M src/kudu/consensus/log.cc M src/kudu/consensus/log_util.cc M src/kudu/fs/block_manager-test.cc M src/kudu/fs/log_block_manager.cc M src/kudu/fs/log_block_manager.h M src/kudu/integration-tests/CMakeLists.txt A src/kudu/integration-tests/disk_reservation-itest.cc M src/kudu/tablet/tablet_peer_mm_ops.cc M src/kudu/util/CMakeLists.txt A src/kudu/util/env_util-test.cc M src/kudu/util/env_util.cc M src/kudu/util/env_util.h M src/kudu/util/scoped_cleanup.h 14 files changed, 613 insertions(+), 76 deletions(-) Approvals: Jean-Daniel Cryans: Looks good to me, but someone else must approve Mike Percy: Verified Adar Dembo: Looks good to me, approved Todd Lipcon: Looks good to me, but someone else must approve -- To view, visit http://gerrit.cloudera.org:8080/3135 To unsubscribe, visit http://gerrit.cloudera.org:8080/settings Gerrit-MessageType: merged Gerrit-Change-Id: Ifd0451d4dbddc1783019a53302de0263080939c7 Gerrit-PatchSet: 10 Gerrit-Project: kudu Gerrit-Branch: master Gerrit-Owner: Mike Percy Gerrit-Reviewer: Adar Dembo Gerrit-Reviewer: Jean-Daniel Cryans Gerrit-Reviewer: Mike Percy Gerrit-Reviewer: Todd Lipcon
[kudu-CR] catalog manager: prevent spurious dirty callbacks from crashing the process
Mike Percy has submitted this change and it was merged. Change subject: catalog_manager: prevent spurious dirty callbacks from crashing the process .. catalog_manager: prevent spurious dirty callbacks from crashing the process The recent change to switch from LocalConsensus to RaftConsensus has led to an increased number of dirty callback calls when using just one master. Some of these calls occur with no term change; the catalog manager treats them as any other calls and reloads the on-disk metadata. In theory that's just unnecessary work, but CheckIsLeaderAndReady() doesn't provide adequate protection for the rest of the master when in this state, so nearly every RPC is allowed in during this time. That's an absolute disaster for correctness: imagine a GetTableLocations() returning only a subset of a table's tablets because the rest were still being loaded from disk. Luckily for us it can also manifest as a crash [1] so we noticed it quickly. I chose to fix it by ignoring these calls within CatalogManager::VisitTablesAndTabletsTask and not SysCatalogTable::SysCatalogStateChanged because the synchronization in the former is more straight forward thanks to the size of worker_pool_. The new test led to a crash 100% of the time without this fix. There's an argument to be made for changing TableInfo::TabletMap's raw pointers to shared pointers thus avoiding this crash altogether, but it's still an incorrect state to be in, so I don't see the value in doing that. While I was here, I snuck a few other changes in: - Remove a lock acquisition from ElectedAsLeaderCb; it did nothing. - Remove old_role_ from SysCatalogTable; it also did nothing. - Narrow the lock acquisition in VisitTablesAndTabletsTask; it only needs to protect the visiting logic. 1. Sample crash output: F0618 05:47:41.795367 9330 ref_counted.cc:74] Check failed: !in_dtor_ *** Check failure stack trace: *** @ 0x7f99fab05f7d google::LogMessage::Fail() at ??:0 @ 0x7f99fab07e7d google::LogMessage::SendToLog() at ??:0 @ 0x7f99fab05ab9 google::LogMessage::Flush() at ??:0 @ 0x7f99fab0891f google::LogMessageFatal::~LogMessageFatal() at ??:0 @ 0x7f99fae8a637 kudu::subtle::RefCountedThreadSafeBase::AddRef() at ??:0 @ 0x7f9a07c486d8 make_scoped_refptr<>() at ??:0 @ 0x7f9a07c2f7f7 kudu::master::TableInfo::GetTabletsInRange() at ??:0 @ 0x7f9a07c2ec35 kudu::master::CatalogManager::GetTableLocations() at ??:0 @ 0x51962e kudu::master::WaitForRunningTabletCount() at ... @ 0x51ce59 kudu::CreateTableStressTest_RestartMasterDuringCreation_Test::TestBody() at ... @ 0x7f99fc7f5b48 testing::internal::HandleExceptionsInMethodIfSupported<>() at ??:0 @ 0x7f99fc7ea012 testing::Test::Run() at ??:0 @ 0x7f99fc7ea158 testing::TestInfo::Run() at ??:0 @ 0x7f99fc7ea235 testing::TestCase::Run() at ??:0 @ 0x7f99fc7ea518 testing::internal::UnitTestImpl::RunAllTests() at ??:0 @ 0x7f99fc7f6058 testing::internal::HandleExceptionsInMethodIfSupported<>() at ??:0 @ 0x7f99fc7ea7fd testing::UnitTest::Run() at ??:0 @ 0x7f9a08520cf6 main at ??:0 @ 0x7f99f97a9d5d __libc_start_main at ??:0 @ 0x43e06d (unknown) at ??:0 Change-Id: I5bc602a51a4914ec13f926bfca555c816a2ee509 Reviewed-on: http://gerrit.cloudera.org:8080/3465 Tested-by: Kudu Jenkins Reviewed-by: Todd Lipcon Reviewed-by: Mike Percy --- M src/kudu/master/catalog_manager.cc M src/kudu/master/catalog_manager.h M src/kudu/master/master-test.cc M src/kudu/master/sys_catalog.cc M src/kudu/master/sys_catalog.h 5 files changed, 82 insertions(+), 48 deletions(-) Approvals: Mike Percy: Looks good to me, approved Todd Lipcon: Looks good to me, but someone else must approve Kudu Jenkins: Verified -- To view, visit http://gerrit.cloudera.org:8080/3465 To unsubscribe, visit http://gerrit.cloudera.org:8080/settings Gerrit-MessageType: merged Gerrit-Change-Id: I5bc602a51a4914ec13f926bfca555c816a2ee509 Gerrit-PatchSet: 4 Gerrit-Project: kudu Gerrit-Branch: master Gerrit-Owner: Adar Dembo Gerrit-Reviewer: Adar Dembo Gerrit-Reviewer: Kudu Jenkins Gerrit-Reviewer: Mike Percy Gerrit-Reviewer: Todd Lipcon
[kudu-CR] catalog manager: prevent spurious dirty callbacks from crashing the process
Mike Percy has posted comments on this change. Change subject: catalog_manager: prevent spurious dirty callbacks from crashing the process .. Patch Set 3: Code-Review+2 -- To view, visit http://gerrit.cloudera.org:8080/3465 To unsubscribe, visit http://gerrit.cloudera.org:8080/settings Gerrit-MessageType: comment Gerrit-Change-Id: I5bc602a51a4914ec13f926bfca555c816a2ee509 Gerrit-PatchSet: 3 Gerrit-Project: kudu Gerrit-Branch: master Gerrit-Owner: Adar Dembo Gerrit-Reviewer: Adar Dembo Gerrit-Reviewer: Kudu Jenkins Gerrit-Reviewer: Mike Percy Gerrit-Reviewer: Todd Lipcon Gerrit-HasComments: No
[kudu-CR] catalog manager: prevent spurious dirty callbacks from crashing the process
Todd Lipcon has posted comments on this change. Change subject: catalog_manager: prevent spurious dirty callbacks from crashing the process .. Patch Set 3: Code-Review+1 -- To view, visit http://gerrit.cloudera.org:8080/3465 To unsubscribe, visit http://gerrit.cloudera.org:8080/settings Gerrit-MessageType: comment Gerrit-Change-Id: I5bc602a51a4914ec13f926bfca555c816a2ee509 Gerrit-PatchSet: 3 Gerrit-Project: kudu Gerrit-Branch: master Gerrit-Owner: Adar Dembo Gerrit-Reviewer: Adar Dembo Gerrit-Reviewer: Kudu Jenkins Gerrit-Reviewer: Mike Percy Gerrit-Reviewer: Todd Lipcon Gerrit-HasComments: No
[kudu-CR] Add required Debian version to installation page
Dan Burkert has posted comments on this change. Change subject: Add required Debian version to installation page .. Patch Set 1: -Code-Review JD, are you ok with pushing this now with your recent deb 8 testing? -- To view, visit http://gerrit.cloudera.org:8080/3389 To unsubscribe, visit http://gerrit.cloudera.org:8080/settings Gerrit-MessageType: comment Gerrit-Change-Id: I65cef323648d7934cb140046cf391d56867eeb56 Gerrit-PatchSet: 1 Gerrit-Project: kudu Gerrit-Branch: master Gerrit-Owner: Dan Burkert Gerrit-Reviewer: Adar Dembo Gerrit-Reviewer: Dan Burkert Gerrit-Reviewer: Jean-Daniel Cryans Gerrit-Reviewer: Kudu Jenkins Gerrit-HasComments: No
[kudu-CR] Allow for reserving disk space for non-Kudu processes
Todd Lipcon has posted comments on this change. Change subject: Allow for reserving disk space for non-Kudu processes .. Patch Set 9: Code-Review+1 -- To view, visit http://gerrit.cloudera.org:8080/3135 To unsubscribe, visit http://gerrit.cloudera.org:8080/settings Gerrit-MessageType: comment Gerrit-Change-Id: Ifd0451d4dbddc1783019a53302de0263080939c7 Gerrit-PatchSet: 9 Gerrit-Project: kudu Gerrit-Branch: master Gerrit-Owner: Mike Percy Gerrit-Reviewer: Adar Dembo Gerrit-Reviewer: Jean-Daniel Cryans Gerrit-Reviewer: Mike Percy Gerrit-Reviewer: Todd Lipcon Gerrit-HasComments: No
[kudu-CR] KUDU-1398 CFile index blocks can store shortest separating prefix
Todd Lipcon has posted comments on this change. Change subject: KUDU-1398 CFile index blocks can store shortest separating prefix .. Patch Set 11: Code-Review+2 -- To view, visit http://gerrit.cloudera.org:8080/3304 To unsubscribe, visit http://gerrit.cloudera.org:8080/settings Gerrit-MessageType: comment Gerrit-Change-Id: I68ae9146fabd4a19b17d103d118d2d60e28bb315 Gerrit-PatchSet: 11 Gerrit-Project: kudu Gerrit-Branch: master Gerrit-Owner: Will Berkeley Gerrit-Reviewer: Kudu Jenkins Gerrit-Reviewer: Todd Lipcon Gerrit-Reviewer: Will Berkeley Gerrit-HasComments: No
[kudu-CR] KUDU-1398 CFile index blocks can store shortest separating prefix
Todd Lipcon has submitted this change and it was merged. Change subject: KUDU-1398 CFile index blocks can store shortest separating prefix .. KUDU-1398 CFile index blocks can store shortest separating prefix (No changes: resubmitting to trigger Jenkins build) This changes the values stored as index keys to be a shortest key between the first key of the data block and the last key of the previous data block. However, this change does not apply to deltafiles, because deltafiles expect to be able to decode an index key into a DeltaKey. The way the change works is illustrated with the example from the JIRA, extended: Block 1: apple, banana, cardamom Block 2: carrot, epazote, fennel Block 3: fig, guava, kiwi Before: ['apple' -> block 1, 'carrot' -> block 2, 'fig' -> block 3] After: ['a' -> block 1, 'carr' -> block 2, 'fi' -> block 3] Change-Id: I68ae9146fabd4a19b17d103d118d2d60e28bb315 Reviewed-on: http://gerrit.cloudera.org:8080/3304 Tested-by: Kudu Jenkins Reviewed-by: Todd Lipcon --- M src/kudu/cfile/binary_dict_block.cc M src/kudu/cfile/binary_dict_block.h M src/kudu/cfile/binary_plain_block.cc M src/kudu/cfile/binary_plain_block.h M src/kudu/cfile/binary_prefix_block.cc M src/kudu/cfile/binary_prefix_block.h M src/kudu/cfile/block_encodings.h M src/kudu/cfile/bloomfile.cc M src/kudu/cfile/bloomfile.h M src/kudu/cfile/bshuf_block.h M src/kudu/cfile/cfile-test.cc M src/kudu/cfile/cfile_util.cc M src/kudu/cfile/cfile_util.h M src/kudu/cfile/cfile_writer.cc M src/kudu/cfile/cfile_writer.h M src/kudu/cfile/gvint_block.cc M src/kudu/cfile/gvint_block.h M src/kudu/cfile/index-test.cc M src/kudu/cfile/index_block.cc M src/kudu/cfile/plain_bitmap_block.h M src/kudu/cfile/plain_block.h M src/kudu/cfile/rle_block.h M src/kudu/tablet/deltafile.cc 23 files changed, 234 insertions(+), 71 deletions(-) Approvals: Todd Lipcon: Looks good to me, approved Kudu Jenkins: Verified -- To view, visit http://gerrit.cloudera.org:8080/3304 To unsubscribe, visit http://gerrit.cloudera.org:8080/settings Gerrit-MessageType: merged Gerrit-Change-Id: I68ae9146fabd4a19b17d103d118d2d60e28bb315 Gerrit-PatchSet: 12 Gerrit-Project: kudu Gerrit-Branch: master Gerrit-Owner: Will Berkeley Gerrit-Reviewer: Kudu Jenkins Gerrit-Reviewer: Todd Lipcon Gerrit-Reviewer: Will Berkeley
[kudu-CR] Allow for reserving disk space for non-Kudu processes
Adar Dembo has posted comments on this change. Change subject: Allow for reserving disk space for non-Kudu processes .. Patch Set 9: > Any particular areas you think I should take a close look at? Nothing specific, I just wanted to make sure that you were satisfied with how Mike addressed your comments. I would say the same thing in any other review (i.e. if two reviewers look at a patch, I think it's good etiquette for the first one who granted a +2 to delay merging until the second reviewer is satisfied as well). -- To view, visit http://gerrit.cloudera.org:8080/3135 To unsubscribe, visit http://gerrit.cloudera.org:8080/settings Gerrit-MessageType: comment Gerrit-Change-Id: Ifd0451d4dbddc1783019a53302de0263080939c7 Gerrit-PatchSet: 9 Gerrit-Project: kudu Gerrit-Branch: master Gerrit-Owner: Mike Percy Gerrit-Reviewer: Adar Dembo Gerrit-Reviewer: Jean-Daniel Cryans Gerrit-Reviewer: Mike Percy Gerrit-Reviewer: Todd Lipcon Gerrit-HasComments: No
[kudu-CR] Allow for reserving disk space for non-Kudu processes
Jean-Daniel Cryans has posted comments on this change. Change subject: Allow for reserving disk space for non-Kudu processes .. Patch Set 9: Code-Review+1 > JD: > > I'm happy to add docs as a follow up. I've actually not written > much documentation so I'll poke around for examples. Drop a section in http://kudu.apache.org/docs/administration.html, can do a follow-up patch since this looks ready to go. > > In terms of non-zero defaults, I'm not sure what kinds of defaults > would be widely applicable, to be honest. I looked at Hadoop, it defaults to 0 (thought it was some other number), so we should just recommend setting that configuration to something in the docs. -- To view, visit http://gerrit.cloudera.org:8080/3135 To unsubscribe, visit http://gerrit.cloudera.org:8080/settings Gerrit-MessageType: comment Gerrit-Change-Id: Ifd0451d4dbddc1783019a53302de0263080939c7 Gerrit-PatchSet: 9 Gerrit-Project: kudu Gerrit-Branch: master Gerrit-Owner: Mike Percy Gerrit-Reviewer: Adar Dembo Gerrit-Reviewer: Jean-Daniel Cryans Gerrit-Reviewer: Mike Percy Gerrit-Reviewer: Todd Lipcon Gerrit-HasComments: No
[kudu-CR] catalog manager: prevent spurious dirty callbacks from crashing the process
Kudu Jenkins has posted comments on this change. Change subject: catalog_manager: prevent spurious dirty callbacks from crashing the process .. Patch Set 3: Build Started http://104.196.14.100/job/kudu-gerrit/1974/ -- To view, visit http://gerrit.cloudera.org:8080/3465 To unsubscribe, visit http://gerrit.cloudera.org:8080/settings Gerrit-MessageType: comment Gerrit-Change-Id: I5bc602a51a4914ec13f926bfca555c816a2ee509 Gerrit-PatchSet: 3 Gerrit-Project: kudu Gerrit-Branch: master Gerrit-Owner: Adar Dembo Gerrit-Reviewer: Adar Dembo Gerrit-Reviewer: Kudu Jenkins Gerrit-Reviewer: Mike Percy Gerrit-Reviewer: Todd Lipcon Gerrit-HasComments: No
[kudu-CR] catalog manager: prevent spurious dirty callbacks from crashing the process
Hello Kudu Jenkins, I'd like you to reexamine a change. Please visit http://gerrit.cloudera.org:8080/3465 to look at the new patch set (#3). Change subject: catalog_manager: prevent spurious dirty callbacks from crashing the process .. catalog_manager: prevent spurious dirty callbacks from crashing the process The recent change to switch from LocalConsensus to RaftConsensus has led to an increased number of dirty callback calls when using just one master. Some of these calls occur with no term change; the catalog manager treats them as any other calls and reloads the on-disk metadata. In theory that's just unnecessary work, but CheckIsLeaderAndReady() doesn't provide adequate protection for the rest of the master when in this state, so nearly every RPC is allowed in during this time. That's an absolute disaster for correctness: imagine a GetTableLocations() returning only a subset of a table's tablets because the rest were still being loaded from disk. Luckily for us it can also manifest as a crash [1] so we noticed it quickly. I chose to fix it by ignoring these calls within CatalogManager::VisitTablesAndTabletsTask and not SysCatalogTable::SysCatalogStateChanged because the synchronization in the former is more straight forward thanks to the size of worker_pool_. The new test led to a crash 100% of the time without this fix. There's an argument to be made for changing TableInfo::TabletMap's raw pointers to shared pointers thus avoiding this crash altogether, but it's still an incorrect state to be in, so I don't see the value in doing that. While I was here, I snuck a few other changes in: - Remove a lock acquisition from ElectedAsLeaderCb; it did nothing. - Remove old_role_ from SysCatalogTable; it also did nothing. - Narrow the lock acquisition in VisitTablesAndTabletsTask; it only needs to protect the visiting logic. 1. Sample crash output: F0618 05:47:41.795367 9330 ref_counted.cc:74] Check failed: !in_dtor_ *** Check failure stack trace: *** @ 0x7f99fab05f7d google::LogMessage::Fail() at ??:0 @ 0x7f99fab07e7d google::LogMessage::SendToLog() at ??:0 @ 0x7f99fab05ab9 google::LogMessage::Flush() at ??:0 @ 0x7f99fab0891f google::LogMessageFatal::~LogMessageFatal() at ??:0 @ 0x7f99fae8a637 kudu::subtle::RefCountedThreadSafeBase::AddRef() at ??:0 @ 0x7f9a07c486d8 make_scoped_refptr<>() at ??:0 @ 0x7f9a07c2f7f7 kudu::master::TableInfo::GetTabletsInRange() at ??:0 @ 0x7f9a07c2ec35 kudu::master::CatalogManager::GetTableLocations() at ??:0 @ 0x51962e kudu::master::WaitForRunningTabletCount() at ... @ 0x51ce59 kudu::CreateTableStressTest_RestartMasterDuringCreation_Test::TestBody() at ... @ 0x7f99fc7f5b48 testing::internal::HandleExceptionsInMethodIfSupported<>() at ??:0 @ 0x7f99fc7ea012 testing::Test::Run() at ??:0 @ 0x7f99fc7ea158 testing::TestInfo::Run() at ??:0 @ 0x7f99fc7ea235 testing::TestCase::Run() at ??:0 @ 0x7f99fc7ea518 testing::internal::UnitTestImpl::RunAllTests() at ??:0 @ 0x7f99fc7f6058 testing::internal::HandleExceptionsInMethodIfSupported<>() at ??:0 @ 0x7f99fc7ea7fd testing::UnitTest::Run() at ??:0 @ 0x7f9a08520cf6 main at ??:0 @ 0x7f99f97a9d5d __libc_start_main at ??:0 @ 0x43e06d (unknown) at ??:0 Change-Id: I5bc602a51a4914ec13f926bfca555c816a2ee509 --- M src/kudu/master/catalog_manager.cc M src/kudu/master/catalog_manager.h M src/kudu/master/master-test.cc M src/kudu/master/sys_catalog.cc M src/kudu/master/sys_catalog.h 5 files changed, 82 insertions(+), 48 deletions(-) git pull ssh://gerrit.cloudera.org:29418/kudu refs/changes/65/3465/3 -- To view, visit http://gerrit.cloudera.org:8080/3465 To unsubscribe, visit http://gerrit.cloudera.org:8080/settings Gerrit-MessageType: newpatchset Gerrit-Change-Id: I5bc602a51a4914ec13f926bfca555c816a2ee509 Gerrit-PatchSet: 3 Gerrit-Project: kudu Gerrit-Branch: master Gerrit-Owner: Adar Dembo Gerrit-Reviewer: Adar Dembo Gerrit-Reviewer: Kudu Jenkins Gerrit-Reviewer: Mike Percy Gerrit-Reviewer: Todd Lipcon
[kudu-CR] catalog manager: prevent spurious dirty callbacks from crashing the process
Adar Dembo has posted comments on this change. Change subject: catalog_manager: prevent spurious dirty callbacks from crashing the process .. Patch Set 2: (2 comments) http://gerrit.cloudera.org:8080/#/c/3465/2/src/kudu/master/catalog_manager.cc File src/kudu/master/catalog_manager.cc: Line 600: std::lock_guard l(state_lock_); > Could you add a DCHECK here that the worker pool has a maximum of 1 threads DCHECK() isn't possible without exposing some ThreadPool innards, but there is already a comment in catalog_manager.h, just above the definition of worker_pool_. I moved it to the constructor because that's where we actually set the number of threads. http://gerrit.cloudera.org:8080/#/c/3465/2/src/kudu/master/master-test.cc File src/kudu/master/master-test.cc: Line 586: static void GetTableLocations(const char* kTableName, > Rename to LoopGetTableLocations? Sure. Renamed GetTableSchema too. -- To view, visit http://gerrit.cloudera.org:8080/3465 To unsubscribe, visit http://gerrit.cloudera.org:8080/settings Gerrit-MessageType: comment Gerrit-Change-Id: I5bc602a51a4914ec13f926bfca555c816a2ee509 Gerrit-PatchSet: 2 Gerrit-Project: kudu Gerrit-Branch: master Gerrit-Owner: Adar Dembo Gerrit-Reviewer: Adar Dembo Gerrit-Reviewer: Kudu Jenkins Gerrit-Reviewer: Mike Percy Gerrit-Reviewer: Todd Lipcon Gerrit-HasComments: Yes
[kudu-CR] Allow for reserving disk space for non-Kudu processes
Todd Lipcon has posted comments on this change. Change subject: Allow for reserving disk space for non-Kudu processes .. Patch Set 9: Any particular areas you think I should take a close look at? Otherwise happy to trust that this has already gotten enough review from Adar -- To view, visit http://gerrit.cloudera.org:8080/3135 To unsubscribe, visit http://gerrit.cloudera.org:8080/settings Gerrit-MessageType: comment Gerrit-Change-Id: Ifd0451d4dbddc1783019a53302de0263080939c7 Gerrit-PatchSet: 9 Gerrit-Project: kudu Gerrit-Branch: master Gerrit-Owner: Mike Percy Gerrit-Reviewer: Adar Dembo Gerrit-Reviewer: Jean-Daniel Cryans Gerrit-Reviewer: Mike Percy Gerrit-Reviewer: Todd Lipcon Gerrit-HasComments: No
[kudu-CR] catalog manager: prevent spurious dirty callbacks from crashing the process
Mike Percy has posted comments on this change. Change subject: catalog_manager: prevent spurious dirty callbacks from crashing the process .. Patch Set 2: (2 comments) http://gerrit.cloudera.org:8080/#/c/3465/2/src/kudu/master/catalog_manager.cc File src/kudu/master/catalog_manager.cc: Line 600: std::lock_guard l(state_lock_); Could you add a DCHECK here that the worker pool has a maximum of 1 threads? If that's hard, how about a comment where the worker_pool_ is constructed that specifies that we rely on that for correctness here. http://gerrit.cloudera.org:8080/#/c/3465/2/src/kudu/master/master-test.cc File src/kudu/master/master-test.cc: Line 586: static void GetTableLocations(const char* kTableName, Rename to LoopGetTableLocations? -- To view, visit http://gerrit.cloudera.org:8080/3465 To unsubscribe, visit http://gerrit.cloudera.org:8080/settings Gerrit-MessageType: comment Gerrit-Change-Id: I5bc602a51a4914ec13f926bfca555c816a2ee509 Gerrit-PatchSet: 2 Gerrit-Project: kudu Gerrit-Branch: master Gerrit-Owner: Adar Dembo Gerrit-Reviewer: Kudu Jenkins Gerrit-Reviewer: Mike Percy Gerrit-Reviewer: Todd Lipcon Gerrit-HasComments: Yes
[kudu-CR] Allow for reserving disk space for non-Kudu processes
Mike Percy has posted comments on this change. Change subject: Allow for reserving disk space for non-Kudu processes .. Patch Set 9: JD: I'm happy to add docs as a follow up. I've actually not written much documentation so I'll poke around for examples. In terms of non-zero defaults, I'm not sure what kinds of defaults would be widely applicable, to be honest. -- To view, visit http://gerrit.cloudera.org:8080/3135 To unsubscribe, visit http://gerrit.cloudera.org:8080/settings Gerrit-MessageType: comment Gerrit-Change-Id: Ifd0451d4dbddc1783019a53302de0263080939c7 Gerrit-PatchSet: 9 Gerrit-Project: kudu Gerrit-Branch: master Gerrit-Owner: Mike Percy Gerrit-Reviewer: Adar Dembo Gerrit-Reviewer: Jean-Daniel Cryans Gerrit-Reviewer: Mike Percy Gerrit-Reviewer: Todd Lipcon Gerrit-HasComments: No
[kudu-CR] KUDU-1398 CFile index blocks can store shortest separating prefix
Kudu Jenkins has posted comments on this change. Change subject: KUDU-1398 CFile index blocks can store shortest separating prefix .. Patch Set 11: Build Started http://104.196.14.100/job/kudu-gerrit/1973/ -- To view, visit http://gerrit.cloudera.org:8080/3304 To unsubscribe, visit http://gerrit.cloudera.org:8080/settings Gerrit-MessageType: comment Gerrit-Change-Id: I68ae9146fabd4a19b17d103d118d2d60e28bb315 Gerrit-PatchSet: 11 Gerrit-Project: kudu Gerrit-Branch: master Gerrit-Owner: Will Berkeley Gerrit-Reviewer: Kudu Jenkins Gerrit-Reviewer: Todd Lipcon Gerrit-Reviewer: Will Berkeley Gerrit-HasComments: No
[kudu-CR] Allow for reserving disk space for non-Kudu processes
Jean-Daniel Cryans has posted comments on this change. Change subject: Allow for reserving disk space for non-Kudu processes .. Patch Set 9: Hey Mike, are you planning on doing a follow-up patch with some documentation? Also, what does it take so that we don't default this to 0? I think it's desirable to have it enabled. -- To view, visit http://gerrit.cloudera.org:8080/3135 To unsubscribe, visit http://gerrit.cloudera.org:8080/settings Gerrit-MessageType: comment Gerrit-Change-Id: Ifd0451d4dbddc1783019a53302de0263080939c7 Gerrit-PatchSet: 9 Gerrit-Project: kudu Gerrit-Branch: master Gerrit-Owner: Mike Percy Gerrit-Reviewer: Adar Dembo Gerrit-Reviewer: Jean-Daniel Cryans Gerrit-Reviewer: Mike Percy Gerrit-Reviewer: Todd Lipcon Gerrit-HasComments: No
[kudu-CR] alter schema transaction: don't crash in ToString() when no timestamp
Adar Dembo has submitted this change and it was merged. Change subject: alter_schema_transaction: don't crash in ToString() when no timestamp .. alter_schema_transaction: don't crash in ToString() when no timestamp Saw this in a local alter_table-test failure. I0623 14:17:50.224598 26418 transaction_tracker.cc:230] Dumping currently running transactions: F0623 14:17:50.224697 26418 transaction.h:206] Check failed: timestamp_ != Timestamp::kInvalidTimestamp *** Check failure stack trace: *** @ 0x7f4bb399146d google::LogMessage::Fail() at ??:0 @ 0x7f4bb39934d3 google::LogMessage::SendToLog() at ??:0 @ 0x7f4bb3990fca google::LogMessage::Flush() at ??:0 @ 0x7f4bb3993f1f google::LogMessageFatal::~LogMessageFatal() at ??:0 @ 0x7f4bb484e3f6 kudu::tablet::TransactionState::timestamp() at ??:0 @ 0x7f4bb489b804 kudu::tablet::AlterSchemaTransactionState::ToString() at ??:0 @ 0x7f4bb489c8f6 kudu::tablet::AlterSchemaTransaction::ToString() at ??:0 @ 0x7f4bb489f1fd kudu::tablet::TransactionDriver::ToStringUnlocked() at ??:0 @ 0x7f4bb489f132 kudu::tablet::TransactionDriver::ToString() at ??:0 @ 0x7f4bb48a6790 kudu::tablet::TransactionTracker::WaitForAllToFinish() at ??:0 @ 0x7f4bb48a612c kudu::tablet::TransactionTracker::WaitForAllToFinish() at ??:0 @ 0x7f4bb4891100 kudu::tablet::TabletPeer::Shutdown() at ??:0 @ 0x7f4bb4bfc993 kudu::tserver::TSTabletManager::Shutdown() at ??:0 @ 0x7f4bb4bd11a2 kudu::tserver::TabletServer::Shutdown() at ??:0 @ 0x7f4bb4b97e71 kudu::tserver::MiniTabletServer::Shutdown() at ??:0 @ 0x7f4bb4b97f8c kudu::tserver::MiniTabletServer::Restart() at ??:0 @ 0x447c10 kudu::AlterTableTest::RestartTabletServer() at /home/adar/Source/kudu/src/kudu/integration-tests/alter_table-test.cc:159 @ 0x4312e3 kudu::AlterTableTest_TestRestartTSDuringAlter_Test::TestBody() at /home/adar/Source/kudu/src/kudu/integration-tests/alter_table-test.cc:394 (discriminator 2) @ 0x7f4bb4ebf0f7 testing::internal::HandleExceptionsInMethodIfSupported<>() at ??:0 @ 0x7f4bb4eb5f92 testing::Test::Run() at ??:0 @ 0x7f4bb4eb60d8 testing::TestInfo::Run() at ??:0 @ 0x7f4bb4eb61b5 testing::TestCase::Run() at ??:0 @ 0x7f4bb4eb6be8 testing::internal::UnitTestImpl::RunAllTests() at ??:0 @ 0x7f4bb4ebf667 testing::internal::HandleExceptionsInMethodIfSupported<>() at ??:0 @ 0x7f4bb4eb628d testing::UnitTest::Run() at ??:0 @ 0x7f4bb4f6e226 RUN_ALL_TESTS() at ??:0 @ 0x7f4bb4f6dc89 main at ??:0 @ 0x7f4bb2a02830 __libc_start_main at ??:0 @ 0x42c689 _start at ??:? @ (nil) (unknown) Change-Id: I702aa5437dec54b93a78c9415d83be0d7d07c364 Reviewed-on: http://gerrit.cloudera.org:8080/3475 Reviewed-by: David Ribeiro Alves Tested-by: Kudu Jenkins --- M src/kudu/tablet/transactions/alter_schema_transaction.cc 1 file changed, 1 insertion(+), 1 deletion(-) Approvals: David Ribeiro Alves: Looks good to me, approved Kudu Jenkins: Verified -- To view, visit http://gerrit.cloudera.org:8080/3475 To unsubscribe, visit http://gerrit.cloudera.org:8080/settings Gerrit-MessageType: merged Gerrit-Change-Id: I702aa5437dec54b93a78c9415d83be0d7d07c364 Gerrit-PatchSet: 2 Gerrit-Project: kudu Gerrit-Branch: master Gerrit-Owner: Adar Dembo Gerrit-Reviewer: Adar Dembo Gerrit-Reviewer: David Ribeiro Alves Gerrit-Reviewer: Kudu Jenkins
[kudu-CR] alter schema transaction: don't crash in ToString() when no timestamp
David Ribeiro Alves has posted comments on this change. Change subject: alter_schema_transaction: don't crash in ToString() when no timestamp .. Patch Set 1: Code-Review+2 -- To view, visit http://gerrit.cloudera.org:8080/3475 To unsubscribe, visit http://gerrit.cloudera.org:8080/settings Gerrit-MessageType: comment Gerrit-Change-Id: I702aa5437dec54b93a78c9415d83be0d7d07c364 Gerrit-PatchSet: 1 Gerrit-Project: kudu Gerrit-Branch: master Gerrit-Owner: Adar Dembo Gerrit-Reviewer: David Ribeiro Alves Gerrit-Reviewer: Kudu Jenkins Gerrit-HasComments: No
[kudu-CR] alter schema transaction: don't crash in ToString() when no timestamp
Kudu Jenkins has posted comments on this change. Change subject: alter_schema_transaction: don't crash in ToString() when no timestamp .. Patch Set 1: Build Started http://104.196.14.100/job/kudu-gerrit/1972/ -- To view, visit http://gerrit.cloudera.org:8080/3475 To unsubscribe, visit http://gerrit.cloudera.org:8080/settings Gerrit-MessageType: comment Gerrit-Change-Id: I702aa5437dec54b93a78c9415d83be0d7d07c364 Gerrit-PatchSet: 1 Gerrit-Project: kudu Gerrit-Branch: master Gerrit-Owner: Adar Dembo Gerrit-Reviewer: David Ribeiro Alves Gerrit-Reviewer: Kudu Jenkins Gerrit-HasComments: No
[kudu-CR] alter schema transaction: don't crash in ToString() when no timestamp
Hello David Ribeiro Alves, I'd like you to do a code review. Please visit http://gerrit.cloudera.org:8080/3475 to review the following change. Change subject: alter_schema_transaction: don't crash in ToString() when no timestamp .. alter_schema_transaction: don't crash in ToString() when no timestamp Saw this in a local alter_table-test failure. I0623 14:17:50.224598 26418 transaction_tracker.cc:230] Dumping currently running transactions: F0623 14:17:50.224697 26418 transaction.h:206] Check failed: timestamp_ != Timestamp::kInvalidTimestamp *** Check failure stack trace: *** @ 0x7f4bb399146d google::LogMessage::Fail() at ??:0 @ 0x7f4bb39934d3 google::LogMessage::SendToLog() at ??:0 @ 0x7f4bb3990fca google::LogMessage::Flush() at ??:0 @ 0x7f4bb3993f1f google::LogMessageFatal::~LogMessageFatal() at ??:0 @ 0x7f4bb484e3f6 kudu::tablet::TransactionState::timestamp() at ??:0 @ 0x7f4bb489b804 kudu::tablet::AlterSchemaTransactionState::ToString() at ??:0 @ 0x7f4bb489c8f6 kudu::tablet::AlterSchemaTransaction::ToString() at ??:0 @ 0x7f4bb489f1fd kudu::tablet::TransactionDriver::ToStringUnlocked() at ??:0 @ 0x7f4bb489f132 kudu::tablet::TransactionDriver::ToString() at ??:0 @ 0x7f4bb48a6790 kudu::tablet::TransactionTracker::WaitForAllToFinish() at ??:0 @ 0x7f4bb48a612c kudu::tablet::TransactionTracker::WaitForAllToFinish() at ??:0 @ 0x7f4bb4891100 kudu::tablet::TabletPeer::Shutdown() at ??:0 @ 0x7f4bb4bfc993 kudu::tserver::TSTabletManager::Shutdown() at ??:0 @ 0x7f4bb4bd11a2 kudu::tserver::TabletServer::Shutdown() at ??:0 @ 0x7f4bb4b97e71 kudu::tserver::MiniTabletServer::Shutdown() at ??:0 @ 0x7f4bb4b97f8c kudu::tserver::MiniTabletServer::Restart() at ??:0 @ 0x447c10 kudu::AlterTableTest::RestartTabletServer() at /home/adar/Source/kudu/src/kudu/integration-tests/alter_table-test.cc:159 @ 0x4312e3 kudu::AlterTableTest_TestRestartTSDuringAlter_Test::TestBody() at /home/adar/Source/kudu/src/kudu/integration-tests/alter_table-test.cc:394 (discriminator 2) @ 0x7f4bb4ebf0f7 testing::internal::HandleExceptionsInMethodIfSupported<>() at ??:0 @ 0x7f4bb4eb5f92 testing::Test::Run() at ??:0 @ 0x7f4bb4eb60d8 testing::TestInfo::Run() at ??:0 @ 0x7f4bb4eb61b5 testing::TestCase::Run() at ??:0 @ 0x7f4bb4eb6be8 testing::internal::UnitTestImpl::RunAllTests() at ??:0 @ 0x7f4bb4ebf667 testing::internal::HandleExceptionsInMethodIfSupported<>() at ??:0 @ 0x7f4bb4eb628d testing::UnitTest::Run() at ??:0 @ 0x7f4bb4f6e226 RUN_ALL_TESTS() at ??:0 @ 0x7f4bb4f6dc89 main at ??:0 @ 0x7f4bb2a02830 __libc_start_main at ??:0 @ 0x42c689 _start at ??:? @ (nil) (unknown) Change-Id: I702aa5437dec54b93a78c9415d83be0d7d07c364 --- M src/kudu/tablet/transactions/alter_schema_transaction.cc 1 file changed, 1 insertion(+), 1 deletion(-) git pull ssh://gerrit.cloudera.org:29418/kudu refs/changes/75/3475/1 -- To view, visit http://gerrit.cloudera.org:8080/3475 To unsubscribe, visit http://gerrit.cloudera.org:8080/settings Gerrit-MessageType: newchange Gerrit-Change-Id: I702aa5437dec54b93a78c9415d83be0d7d07c364 Gerrit-PatchSet: 1 Gerrit-Project: kudu Gerrit-Branch: master Gerrit-Owner: Adar Dembo Gerrit-Reviewer: David Ribeiro Alves
[kudu-CR] Release notes for 0.9.1
Todd Lipcon has submitted this change and it was merged. Change subject: Release notes for 0.9.1 .. Release notes for 0.9.1 Change-Id: I797980d105e0e1000420aedf58e094f9505053c6 Reviewed-on: http://gerrit.cloudera.org:8080/3463 Reviewed-by: Misty Stanley-Jones Reviewed-by: Jean-Daniel Cryans Tested-by: Todd Lipcon Reviewed-on: http://gerrit.cloudera.org:8080/3470 Tested-by: Kudu Jenkins --- M docs/installation.adoc M docs/release_notes.adoc 2 files changed, 34 insertions(+), 3 deletions(-) Approvals: Jean-Daniel Cryans: Looks good to me, approved Kudu Jenkins: Verified -- To view, visit http://gerrit.cloudera.org:8080/3470 To unsubscribe, visit http://gerrit.cloudera.org:8080/settings Gerrit-MessageType: merged Gerrit-Change-Id: I797980d105e0e1000420aedf58e094f9505053c6 Gerrit-PatchSet: 2 Gerrit-Project: kudu Gerrit-Branch: master Gerrit-Owner: Todd Lipcon Gerrit-Reviewer: Jean-Daniel Cryans Gerrit-Reviewer: Kudu Jenkins Gerrit-Reviewer: Misty Stanley-Jones Gerrit-Reviewer: Todd Lipcon
[kudu-CR] Allow for reserving disk space for non-Kudu processes
Mike Percy has posted comments on this change. Change subject: Allow for reserving disk space for non-Kudu processes .. Patch Set 9: Verified+1 Overriding apparent Java test flake. Filed https://issues.apache.org/jira/browse/KUDU-1496 about it. Looked at the log and nothing seemed it could potentially be related to this patch. -- To view, visit http://gerrit.cloudera.org:8080/3135 To unsubscribe, visit http://gerrit.cloudera.org:8080/settings Gerrit-MessageType: comment Gerrit-Change-Id: Ifd0451d4dbddc1783019a53302de0263080939c7 Gerrit-PatchSet: 9 Gerrit-Project: kudu Gerrit-Branch: master Gerrit-Owner: Mike Percy Gerrit-Reviewer: Adar Dembo Gerrit-Reviewer: Mike Percy Gerrit-Reviewer: Todd Lipcon Gerrit-HasComments: No
[kudu-CR] Allow for reserving disk space for non-Kudu processes
Adar Dembo has posted comments on this change. Change subject: Allow for reserving disk space for non-Kudu processes .. Patch Set 9: Code-Review+2 Leaving unmerged in case Todd wants to rereview. -- To view, visit http://gerrit.cloudera.org:8080/3135 To unsubscribe, visit http://gerrit.cloudera.org:8080/settings Gerrit-MessageType: comment Gerrit-Change-Id: Ifd0451d4dbddc1783019a53302de0263080939c7 Gerrit-PatchSet: 9 Gerrit-Project: kudu Gerrit-Branch: master Gerrit-Owner: Mike Percy Gerrit-Reviewer: Adar Dembo Gerrit-Reviewer: Mike Percy Gerrit-Reviewer: Todd Lipcon Gerrit-HasComments: No
[kudu-CR](branch-0.9.x) Bump version to 0.9.1
Todd Lipcon has submitted this change and it was merged. Change subject: Bump version to 0.9.1 .. Bump version to 0.9.1 Change-Id: Ied911c99b45f9bdae20d14f0ecbe2ae9010f1476 Reviewed-on: http://gerrit.cloudera.org:8080/3473 Reviewed-by: Jean-Daniel Cryans Tested-by: Todd Lipcon --- M java/interface-annotations/pom.xml M java/kudu-client-tools/pom.xml M java/kudu-client/pom.xml M java/kudu-csd/pom.xml M java/kudu-flume-sink/pom.xml M java/kudu-mapreduce/pom.xml M java/kudu-spark/pom.xml M java/pom.xml M version.txt 9 files changed, 10 insertions(+), 10 deletions(-) Approvals: Jean-Daniel Cryans: Looks good to me, approved Todd Lipcon: Verified -- To view, visit http://gerrit.cloudera.org:8080/3473 To unsubscribe, visit http://gerrit.cloudera.org:8080/settings Gerrit-MessageType: merged Gerrit-Change-Id: Ied911c99b45f9bdae20d14f0ecbe2ae9010f1476 Gerrit-PatchSet: 2 Gerrit-Project: kudu Gerrit-Branch: branch-0.9.x Gerrit-Owner: Todd Lipcon Gerrit-Reviewer: Jean-Daniel Cryans Gerrit-Reviewer: Kudu Jenkins Gerrit-Reviewer: Todd Lipcon
[kudu-CR](branch-0.9.x) Bump version to 0.9.1
Todd Lipcon has posted comments on this change. Change subject: Bump version to 0.9.1 .. Patch Set 1: Verified+1 -- To view, visit http://gerrit.cloudera.org:8080/3473 To unsubscribe, visit http://gerrit.cloudera.org:8080/settings Gerrit-MessageType: comment Gerrit-Change-Id: Ied911c99b45f9bdae20d14f0ecbe2ae9010f1476 Gerrit-PatchSet: 1 Gerrit-Project: kudu Gerrit-Branch: branch-0.9.x Gerrit-Owner: Todd Lipcon Gerrit-Reviewer: Jean-Daniel Cryans Gerrit-Reviewer: Kudu Jenkins Gerrit-Reviewer: Todd Lipcon Gerrit-HasComments: No
[kudu-CR](branch-0.9.x) Bump version to 0.9.1
Jean-Daniel Cryans has posted comments on this change. Change subject: Bump version to 0.9.1 .. Patch Set 1: Code-Review+2 -- To view, visit http://gerrit.cloudera.org:8080/3473 To unsubscribe, visit http://gerrit.cloudera.org:8080/settings Gerrit-MessageType: comment Gerrit-Change-Id: Ied911c99b45f9bdae20d14f0ecbe2ae9010f1476 Gerrit-PatchSet: 1 Gerrit-Project: kudu Gerrit-Branch: branch-0.9.x Gerrit-Owner: Todd Lipcon Gerrit-Reviewer: Jean-Daniel Cryans Gerrit-Reviewer: Kudu Jenkins Gerrit-HasComments: No
[kudu-CR](branch-0.9.x) Bump version to 0.9.1
Kudu Jenkins has posted comments on this change. Change subject: Bump version to 0.9.1 .. Patch Set 1: Build Started http://104.196.14.100/job/kudu-gerrit/1971/ -- To view, visit http://gerrit.cloudera.org:8080/3473 To unsubscribe, visit http://gerrit.cloudera.org:8080/settings Gerrit-MessageType: comment Gerrit-Change-Id: Ied911c99b45f9bdae20d14f0ecbe2ae9010f1476 Gerrit-PatchSet: 1 Gerrit-Project: kudu Gerrit-Branch: branch-0.9.x Gerrit-Owner: Todd Lipcon Gerrit-Reviewer: Jean-Daniel Cryans Gerrit-Reviewer: Kudu Jenkins Gerrit-HasComments: No
[kudu-CR](branch-0.9.x) Bump version to 0.9.1
Hello Jean-Daniel Cryans, I'd like you to do a code review. Please visit http://gerrit.cloudera.org:8080/3473 to review the following change. Change subject: Bump version to 0.9.1 .. Bump version to 0.9.1 Change-Id: Ied911c99b45f9bdae20d14f0ecbe2ae9010f1476 --- M java/interface-annotations/pom.xml M java/kudu-client-tools/pom.xml M java/kudu-client/pom.xml M java/kudu-csd/pom.xml M java/kudu-flume-sink/pom.xml M java/kudu-mapreduce/pom.xml M java/kudu-spark/pom.xml M java/pom.xml M version.txt 9 files changed, 10 insertions(+), 10 deletions(-) git pull ssh://gerrit.cloudera.org:29418/kudu refs/changes/73/3473/1 -- To view, visit http://gerrit.cloudera.org:8080/3473 To unsubscribe, visit http://gerrit.cloudera.org:8080/settings Gerrit-MessageType: newchange Gerrit-Change-Id: Ied911c99b45f9bdae20d14f0ecbe2ae9010f1476 Gerrit-PatchSet: 1 Gerrit-Project: kudu Gerrit-Branch: branch-0.9.x Gerrit-Owner: Todd Lipcon Gerrit-Reviewer: Jean-Daniel Cryans
[kudu-CR] Release notes for 0.9.1
Jean-Daniel Cryans has posted comments on this change. Change subject: Release notes for 0.9.1 .. Patch Set 1: Code-Review+2 -- To view, visit http://gerrit.cloudera.org:8080/3470 To unsubscribe, visit http://gerrit.cloudera.org:8080/settings Gerrit-MessageType: comment Gerrit-Change-Id: I797980d105e0e1000420aedf58e094f9505053c6 Gerrit-PatchSet: 1 Gerrit-Project: kudu Gerrit-Branch: master Gerrit-Owner: Todd Lipcon Gerrit-Reviewer: Jean-Daniel Cryans Gerrit-Reviewer: Kudu Jenkins Gerrit-Reviewer: Misty Stanley-Jones Gerrit-HasComments: No
[kudu-CR] Allow for reserving disk space for non-Kudu processes
Kudu Jenkins has posted comments on this change. Change subject: Allow for reserving disk space for non-Kudu processes .. Patch Set 9: Build Started http://104.196.14.100/job/kudu-gerrit/1970/ -- To view, visit http://gerrit.cloudera.org:8080/3135 To unsubscribe, visit http://gerrit.cloudera.org:8080/settings Gerrit-MessageType: comment Gerrit-Change-Id: Ifd0451d4dbddc1783019a53302de0263080939c7 Gerrit-PatchSet: 9 Gerrit-Project: kudu Gerrit-Branch: master Gerrit-Owner: Mike Percy Gerrit-Reviewer: Adar Dembo Gerrit-Reviewer: Kudu Jenkins Gerrit-Reviewer: Mike Percy Gerrit-Reviewer: Todd Lipcon Gerrit-HasComments: No
[kudu-CR] Allow for reserving disk space for non-Kudu processes
Hello Kudu Jenkins, I'd like you to reexamine a change. Please visit http://gerrit.cloudera.org:8080/3135 to look at the new patch set (#9). Change subject: Allow for reserving disk space for non-Kudu processes .. Allow for reserving disk space for non-Kudu processes Adds gflags to reserve disk space such that Kudu will not use more than specified. Hadoop calls this functionality "du.reserved". If a WAL preallocation is attempted while the log disk is past its reservation limit, the log write will fail. As a result, RaftConsensus will cause the process to crash. The log block manager will use non-full disks if possible until all of the disks are full. If a flush or compaction is attempted when all disks are beyond their configured capacity then the LogBlockManager will return an error. As a result, the maintenance manager task will cause the process to crash. This initial implementation provides a "best effort" approach. Disk space checks are only done at preallocation time, and if writes continue beyond the preallocated point (for both a WAL segment and a data block) those writes will not be prevented. This makes it easier to provide a "friendly" option where the block manager will divert new writes to non-full disks, avoiding a hard crash when only one disk is past its reservation limit. In the future, we may want to add "hard" and "soft" limits, such that going beyond the soft limit will do what we do today, and going beyond the hard limit (say, by writing a very large data block past its preallocation point) will result in a crash. This patch includes: * Unit tests. * End-to-end test for flushing / compaction falling back to non-full disks due to disk space backpressure and finally crashing when there is no space left in any data dir. * End-to-end test for writes failing due to WAL disk space backpressure, causing a crash. Change-Id: Ifd0451d4dbddc1783019a53302de0263080939c7 --- M src/kudu/consensus/log-test.cc M src/kudu/consensus/log.cc M src/kudu/consensus/log_util.cc M src/kudu/fs/block_manager-test.cc M src/kudu/fs/log_block_manager.cc M src/kudu/fs/log_block_manager.h M src/kudu/integration-tests/CMakeLists.txt A src/kudu/integration-tests/disk_reservation-itest.cc M src/kudu/tablet/tablet_peer_mm_ops.cc M src/kudu/util/CMakeLists.txt A src/kudu/util/env_util-test.cc M src/kudu/util/env_util.cc M src/kudu/util/env_util.h M src/kudu/util/scoped_cleanup.h 14 files changed, 613 insertions(+), 76 deletions(-) git pull ssh://gerrit.cloudera.org:29418/kudu refs/changes/35/3135/9 -- To view, visit http://gerrit.cloudera.org:8080/3135 To unsubscribe, visit http://gerrit.cloudera.org:8080/settings Gerrit-MessageType: newpatchset Gerrit-Change-Id: Ifd0451d4dbddc1783019a53302de0263080939c7 Gerrit-PatchSet: 9 Gerrit-Project: kudu Gerrit-Branch: master Gerrit-Owner: Mike Percy Gerrit-Reviewer: Adar Dembo Gerrit-Reviewer: Kudu Jenkins Gerrit-Reviewer: Mike Percy Gerrit-Reviewer: Todd Lipcon
[kudu-CR] Allow for reserving disk space for non-Kudu processes
Mike Percy has posted comments on this change. Change subject: Allow for reserving disk space for non-Kudu processes .. Patch Set 5: (5 comments) http://gerrit.cloudera.org:8080/#/c/3135/8/src/kudu/fs/block_manager-test.cc File src/kudu/fs/block_manager-test.cc: PS8, Line 968: : #endif // defined(__linux__) > Do we still need this sleep now that we're using a FINE clock? No. I thought I removed it... thanks for the catch. http://gerrit.cloudera.org:8080/#/c/3135/5/src/kudu/fs/log_block_manager.cc File src/kudu/fs/log_block_manager.cc: Line 1292: FLAGS_log_container_preallocate_bytes, > I see your point (regarding soft vs. hard limits). So let's proceed with wh Done http://gerrit.cloudera.org:8080/#/c/3135/8/src/kudu/fs/log_block_manager.cc File src/kudu/fs/log_block_manager.cc: PS8, Line 1234: : InsertOrDie(&full_root_indexes > Nit: it's not using indexes anymore. Done Line 1240: // In case one or more root paths have hit their reserved space limit, we > Why not InsertOrDie() here and below? I don't think root_paths_ can have du You're right, it's safe to do now, after my last refactor. PS8, Line 1302: : } : RETURN_NOT_OK(s); // Catch IOErrors, etc. : RETURN_NOT_OK(container->Preallocate(FLAGS_lo > Nit: I think getting the gist of the entire line would be easier if you use Done -- To view, visit http://gerrit.cloudera.org:8080/3135 To unsubscribe, visit http://gerrit.cloudera.org:8080/settings Gerrit-MessageType: comment Gerrit-Change-Id: Ifd0451d4dbddc1783019a53302de0263080939c7 Gerrit-PatchSet: 5 Gerrit-Project: kudu Gerrit-Branch: master Gerrit-Owner: Mike Percy Gerrit-Reviewer: Adar Dembo Gerrit-Reviewer: Kudu Jenkins Gerrit-Reviewer: Mike Percy Gerrit-Reviewer: Todd Lipcon Gerrit-HasComments: Yes
[kudu-CR] lock manager: fix TSAN error in LockTable::Resize()
Todd Lipcon has submitted this change and it was merged. Change subject: lock_manager: fix TSAN error in LockTable::Resize() .. lock_manager: fix TSAN error in LockTable::Resize() This function was doing a non-atomic read of item_count_, which is decremented outside the lock by Release(). I looped all_types-itest in TSAN which was previously quite flaky and 50/50 passed: http://dist-test.cloudera.org/job?job_id=todd.1466706366.9004 Change-Id: Id9d57c2b36bfa794cbd786641e456ad748c60941 Reviewed-on: http://gerrit.cloudera.org:8080/3469 Reviewed-by: Adar Dembo Tested-by: Kudu Jenkins --- M src/kudu/tablet/lock_manager.cc 1 file changed, 1 insertion(+), 1 deletion(-) Approvals: Adar Dembo: Looks good to me, approved Kudu Jenkins: Verified -- To view, visit http://gerrit.cloudera.org:8080/3469 To unsubscribe, visit http://gerrit.cloudera.org:8080/settings Gerrit-MessageType: merged Gerrit-Change-Id: Id9d57c2b36bfa794cbd786641e456ad748c60941 Gerrit-PatchSet: 2 Gerrit-Project: kudu Gerrit-Branch: master Gerrit-Owner: Todd Lipcon Gerrit-Reviewer: Adar Dembo Gerrit-Reviewer: Kudu Jenkins Gerrit-Reviewer: Todd Lipcon
[kudu-CR] lock manager: fix TSAN error in LockTable::Resize()
Kudu Jenkins has posted comments on this change. Change subject: lock_manager: fix TSAN error in LockTable::Resize() .. Patch Set 1: -Verified Build Started http://104.196.14.100/job/kudu-gerrit/1969/ -- To view, visit http://gerrit.cloudera.org:8080/3469 To unsubscribe, visit http://gerrit.cloudera.org:8080/settings Gerrit-MessageType: comment Gerrit-Change-Id: Id9d57c2b36bfa794cbd786641e456ad748c60941 Gerrit-PatchSet: 1 Gerrit-Project: kudu Gerrit-Branch: master Gerrit-Owner: Todd Lipcon Gerrit-Reviewer: Adar Dembo Gerrit-Reviewer: Kudu Jenkins Gerrit-Reviewer: Todd Lipcon Gerrit-HasComments: No
[kudu-CR] lock manager: fix TSAN error in LockTable::Resize()
Todd Lipcon has posted comments on this change. Change subject: lock_manager: fix TSAN error in LockTable::Resize() .. Patch Set 1: We had a suppression on ScopedRowLock::Release originally (since we first introduced TSAN a couple years ago) -- To view, visit http://gerrit.cloudera.org:8080/3469 To unsubscribe, visit http://gerrit.cloudera.org:8080/settings Gerrit-MessageType: comment Gerrit-Change-Id: Id9d57c2b36bfa794cbd786641e456ad748c60941 Gerrit-PatchSet: 1 Gerrit-Project: kudu Gerrit-Branch: master Gerrit-Owner: Todd Lipcon Gerrit-Reviewer: Adar Dembo Gerrit-Reviewer: Kudu Jenkins Gerrit-Reviewer: Todd Lipcon Gerrit-HasComments: No
[kudu-CR] lock manager: fix TSAN error in LockTable::Resize()
Adar Dembo has posted comments on this change. Change subject: lock_manager: fix TSAN error in LockTable::Resize() .. Patch Set 1: Code-Review+2 Which annotation was hiding this previously? -- To view, visit http://gerrit.cloudera.org:8080/3469 To unsubscribe, visit http://gerrit.cloudera.org:8080/settings Gerrit-MessageType: comment Gerrit-Change-Id: Id9d57c2b36bfa794cbd786641e456ad748c60941 Gerrit-PatchSet: 1 Gerrit-Project: kudu Gerrit-Branch: master Gerrit-Owner: Todd Lipcon Gerrit-Reviewer: Adar Dembo Gerrit-Reviewer: Kudu Jenkins Gerrit-HasComments: No
[kudu-CR] Release notes for 0.9.1
Hello Jean-Daniel Cryans, Misty Stanley-Jones, I'd like you to do a code review. Please visit http://gerrit.cloudera.org:8080/3470 to review the following change. Change subject: Release notes for 0.9.1 .. Release notes for 0.9.1 Change-Id: I797980d105e0e1000420aedf58e094f9505053c6 Reviewed-on: http://gerrit.cloudera.org:8080/3463 Reviewed-by: Misty Stanley-Jones Reviewed-by: Jean-Daniel Cryans Tested-by: Todd Lipcon --- M docs/installation.adoc M docs/release_notes.adoc 2 files changed, 34 insertions(+), 3 deletions(-) git pull ssh://gerrit.cloudera.org:29418/kudu refs/changes/70/3470/1 -- To view, visit http://gerrit.cloudera.org:8080/3470 To unsubscribe, visit http://gerrit.cloudera.org:8080/settings Gerrit-MessageType: newchange Gerrit-Change-Id: I797980d105e0e1000420aedf58e094f9505053c6 Gerrit-PatchSet: 1 Gerrit-Project: kudu Gerrit-Branch: master Gerrit-Owner: Todd Lipcon Gerrit-Reviewer: Jean-Daniel Cryans Gerrit-Reviewer: Misty Stanley-Jones
[kudu-CR] Release notes for 0.9.1
Kudu Jenkins has posted comments on this change. Change subject: Release notes for 0.9.1 .. Patch Set 1: Build Started http://104.196.14.100/job/kudu-gerrit/1968/ -- To view, visit http://gerrit.cloudera.org:8080/3470 To unsubscribe, visit http://gerrit.cloudera.org:8080/settings Gerrit-MessageType: comment Gerrit-Change-Id: I797980d105e0e1000420aedf58e094f9505053c6 Gerrit-PatchSet: 1 Gerrit-Project: kudu Gerrit-Branch: master Gerrit-Owner: Todd Lipcon Gerrit-Reviewer: Jean-Daniel Cryans Gerrit-Reviewer: Kudu Jenkins Gerrit-Reviewer: Misty Stanley-Jones Gerrit-HasComments: No
[kudu-CR] lock manager: fix TSAN error in LockTable::Resize()
Kudu Jenkins has posted comments on this change. Change subject: lock_manager: fix TSAN error in LockTable::Resize() .. Patch Set 1: Build Started http://104.196.14.100/job/kudu-gerrit/1967/ -- To view, visit http://gerrit.cloudera.org:8080/3469 To unsubscribe, visit http://gerrit.cloudera.org:8080/settings Gerrit-MessageType: comment Gerrit-Change-Id: Id9d57c2b36bfa794cbd786641e456ad748c60941 Gerrit-PatchSet: 1 Gerrit-Project: kudu Gerrit-Branch: master Gerrit-Owner: Todd Lipcon Gerrit-Reviewer: Kudu Jenkins Gerrit-HasComments: No
[kudu-CR] lock manager: fix TSAN error in LockTable::Resize()
Todd Lipcon has uploaded a new change for review. http://gerrit.cloudera.org:8080/3469 Change subject: lock_manager: fix TSAN error in LockTable::Resize() .. lock_manager: fix TSAN error in LockTable::Resize() This function was doing a non-atomic read of item_count_, which is decremented outside the lock by Release(). I looped all_types-itest in TSAN which was previously quite flaky and 50/50 passed: http://dist-test.cloudera.org/job?job_id=todd.1466706366.9004 Change-Id: Id9d57c2b36bfa794cbd786641e456ad748c60941 --- M src/kudu/tablet/lock_manager.cc 1 file changed, 1 insertion(+), 1 deletion(-) git pull ssh://gerrit.cloudera.org:29418/kudu refs/changes/69/3469/1 -- To view, visit http://gerrit.cloudera.org:8080/3469 To unsubscribe, visit http://gerrit.cloudera.org:8080/settings Gerrit-MessageType: newchange Gerrit-Change-Id: Id9d57c2b36bfa794cbd786641e456ad748c60941 Gerrit-PatchSet: 1 Gerrit-Project: kudu Gerrit-Branch: master Gerrit-Owner: Todd Lipcon
[kudu-CR](branch-0.9.x) Release notes for 0.9.1
Todd Lipcon has submitted this change and it was merged. Change subject: Release notes for 0.9.1 .. Release notes for 0.9.1 Change-Id: I797980d105e0e1000420aedf58e094f9505053c6 Reviewed-on: http://gerrit.cloudera.org:8080/3463 Reviewed-by: Misty Stanley-Jones Reviewed-by: Jean-Daniel Cryans Tested-by: Todd Lipcon --- M docs/installation.adoc M docs/release_notes.adoc 2 files changed, 34 insertions(+), 3 deletions(-) Approvals: Jean-Daniel Cryans: Looks good to me, approved Misty Stanley-Jones: Looks good to me, but someone else must approve Todd Lipcon: Verified -- To view, visit http://gerrit.cloudera.org:8080/3463 To unsubscribe, visit http://gerrit.cloudera.org:8080/settings Gerrit-MessageType: merged Gerrit-Change-Id: I797980d105e0e1000420aedf58e094f9505053c6 Gerrit-PatchSet: 4 Gerrit-Project: kudu Gerrit-Branch: branch-0.9.x Gerrit-Owner: Todd Lipcon Gerrit-Reviewer: Jean-Daniel Cryans Gerrit-Reviewer: Kudu Jenkins Gerrit-Reviewer: Misty Stanley-Jones Gerrit-Reviewer: Todd Lipcon
[kudu-CR](branch-0.9.x) Release notes for 0.9.1
Todd Lipcon has posted comments on this change. Change subject: Release notes for 0.9.1 .. Patch Set 3: Verified+1 -- To view, visit http://gerrit.cloudera.org:8080/3463 To unsubscribe, visit http://gerrit.cloudera.org:8080/settings Gerrit-MessageType: comment Gerrit-Change-Id: I797980d105e0e1000420aedf58e094f9505053c6 Gerrit-PatchSet: 3 Gerrit-Project: kudu Gerrit-Branch: branch-0.9.x Gerrit-Owner: Todd Lipcon Gerrit-Reviewer: Jean-Daniel Cryans Gerrit-Reviewer: Kudu Jenkins Gerrit-Reviewer: Misty Stanley-Jones Gerrit-Reviewer: Todd Lipcon Gerrit-HasComments: No
[kudu-CR](branch-0.9.x) Release notes for 0.9.1
Jean-Daniel Cryans has posted comments on this change. Change subject: Release notes for 0.9.1 .. Patch Set 3: Code-Review+2 -- To view, visit http://gerrit.cloudera.org:8080/3463 To unsubscribe, visit http://gerrit.cloudera.org:8080/settings Gerrit-MessageType: comment Gerrit-Change-Id: I797980d105e0e1000420aedf58e094f9505053c6 Gerrit-PatchSet: 3 Gerrit-Project: kudu Gerrit-Branch: branch-0.9.x Gerrit-Owner: Todd Lipcon Gerrit-Reviewer: Jean-Daniel Cryans Gerrit-Reviewer: Kudu Jenkins Gerrit-Reviewer: Misty Stanley-Jones Gerrit-Reviewer: Todd Lipcon Gerrit-HasComments: No
[kudu-CR](branch-0.9.x) Release notes for 0.9.1
Misty Stanley-Jones has posted comments on this change. Change subject: Release notes for 0.9.1 .. Patch Set 3: Code-Review+1 -- To view, visit http://gerrit.cloudera.org:8080/3463 To unsubscribe, visit http://gerrit.cloudera.org:8080/settings Gerrit-MessageType: comment Gerrit-Change-Id: I797980d105e0e1000420aedf58e094f9505053c6 Gerrit-PatchSet: 3 Gerrit-Project: kudu Gerrit-Branch: branch-0.9.x Gerrit-Owner: Todd Lipcon Gerrit-Reviewer: Jean-Daniel Cryans Gerrit-Reviewer: Kudu Jenkins Gerrit-Reviewer: Misty Stanley-Jones Gerrit-Reviewer: Todd Lipcon Gerrit-HasComments: No
[kudu-CR](branch-0.9.x) Release notes for 0.9.1
Hello Kudu Jenkins, I'd like you to reexamine a change. Please visit http://gerrit.cloudera.org:8080/3463 to look at the new patch set (#3). Change subject: Release notes for 0.9.1 .. Release notes for 0.9.1 Change-Id: I797980d105e0e1000420aedf58e094f9505053c6 --- M docs/installation.adoc M docs/release_notes.adoc 2 files changed, 34 insertions(+), 3 deletions(-) git pull ssh://gerrit.cloudera.org:29418/kudu refs/changes/63/3463/3 -- To view, visit http://gerrit.cloudera.org:8080/3463 To unsubscribe, visit http://gerrit.cloudera.org:8080/settings Gerrit-MessageType: newpatchset Gerrit-Change-Id: I797980d105e0e1000420aedf58e094f9505053c6 Gerrit-PatchSet: 3 Gerrit-Project: kudu Gerrit-Branch: branch-0.9.x Gerrit-Owner: Todd Lipcon Gerrit-Reviewer: Jean-Daniel Cryans Gerrit-Reviewer: Kudu Jenkins Gerrit-Reviewer: Misty Stanley-Jones Gerrit-Reviewer: Todd Lipcon
[kudu-CR](branch-0.9.x) Release notes for 0.9.1
Kudu Jenkins has posted comments on this change. Change subject: Release notes for 0.9.1 .. Patch Set 3: Build Started http://104.196.14.100/job/kudu-gerrit/1966/ -- To view, visit http://gerrit.cloudera.org:8080/3463 To unsubscribe, visit http://gerrit.cloudera.org:8080/settings Gerrit-MessageType: comment Gerrit-Change-Id: I797980d105e0e1000420aedf58e094f9505053c6 Gerrit-PatchSet: 3 Gerrit-Project: kudu Gerrit-Branch: branch-0.9.x Gerrit-Owner: Todd Lipcon Gerrit-Reviewer: Jean-Daniel Cryans Gerrit-Reviewer: Kudu Jenkins Gerrit-Reviewer: Misty Stanley-Jones Gerrit-Reviewer: Todd Lipcon Gerrit-HasComments: No
[kudu-CR](branch-0.9.x) Release notes for 0.9.1
Misty Stanley-Jones has posted comments on this change. Change subject: Release notes for 0.9.1 .. Patch Set 1: (1 comment) http://gerrit.cloudera.org:8080/#/c/3463/1/docs/installation.adoc File docs/installation.adoc: Line 639: The same procedure may be used to upgrade from 0.9.0 to 0.9.1. Change this to active voice: To upgrade from 0.9.0 to 0.9.1, use the same procedure as the upgrade from 0.8.0 to 0.9.0. (or something like that) -- To view, visit http://gerrit.cloudera.org:8080/3463 To unsubscribe, visit http://gerrit.cloudera.org:8080/settings Gerrit-MessageType: comment Gerrit-Change-Id: I797980d105e0e1000420aedf58e094f9505053c6 Gerrit-PatchSet: 1 Gerrit-Project: kudu Gerrit-Branch: branch-0.9.x Gerrit-Owner: Todd Lipcon Gerrit-Reviewer: Jean-Daniel Cryans Gerrit-Reviewer: Kudu Jenkins Gerrit-Reviewer: Misty Stanley-Jones Gerrit-Reviewer: Todd Lipcon Gerrit-HasComments: Yes
[kudu-CR] Allow for reserving disk space for non-Kudu processes
Adar Dembo has posted comments on this change. Change subject: Allow for reserving disk space for non-Kudu processes .. Patch Set 8: (6 comments) http://gerrit.cloudera.org:8080/#/c/3135/8/src/kudu/fs/block_manager-test.cc File src/kudu/fs/block_manager-test.cc: PS8, Line 968: // Let the disk-full cache expire. : SleepFor(MonoDelta::FromMilliseconds(100)); Do we still need this sleep now that we're using a FINE clock? http://gerrit.cloudera.org:8080/#/c/3135/5/src/kudu/fs/log_block_manager.cc File src/kudu/fs/log_block_manager.cc: Line 1292: } > Yes, preallocation sounds a little bit broken. I see your point (regarding soft vs. hard limits). So let's proceed with what you have, provided: 1. Add a note to the help for data_dirs_reserved_bytes explaining that it'll only work when log_container_preallocate_bytes is non-zero. 2. Add a comment here explaining the "preallocation size as number of bytes we think we're about to consume" inaccuracy. http://gerrit.cloudera.org:8080/#/c/3135/8/src/kudu/fs/log_block_manager.cc File src/kudu/fs/log_block_manager.cc: PS8, Line 1234: Initialize the : // indexes from the FullDiskCache. Nit: it's not using indexes anymore. Line 1240: full_root_paths.insert(root_paths_[i]); Why not InsertOrDie() here and below? I don't think root_paths_ can have duplicates, can it? PS8, Line 1302: "Log block manager: Insufficient disk space under path " :<< container->root_path() << ": Creation of new data blocks under this path " :<< "can be retried after " << FLAGS_log_block_manager_full_disk_cache_seconds :<< " seconds: " << s.ToString(); Nit: I think getting the gist of the entire line would be easier if you used Substitute() to build it. http://gerrit.cloudera.org:8080/#/c/3135/5/src/kudu/util/env_util.cc File src/kudu/util/env_util.cc: Line 46: TAG_FLAG(disk_reserved_prefixes_with_bytes_free_for_testing, unsafe); > I'm changing them in the integration tests. Ah, didn't notice that. -- To view, visit http://gerrit.cloudera.org:8080/3135 To unsubscribe, visit http://gerrit.cloudera.org:8080/settings Gerrit-MessageType: comment Gerrit-Change-Id: Ifd0451d4dbddc1783019a53302de0263080939c7 Gerrit-PatchSet: 8 Gerrit-Project: kudu Gerrit-Branch: master Gerrit-Owner: Mike Percy Gerrit-Reviewer: Adar Dembo Gerrit-Reviewer: Kudu Jenkins Gerrit-Reviewer: Mike Percy Gerrit-Reviewer: Todd Lipcon Gerrit-HasComments: Yes
[kudu-CR](branch-0.9.x) Release notes for 0.9.1
Hello Kudu Jenkins, I'd like you to reexamine a change. Please visit http://gerrit.cloudera.org:8080/3463 to look at the new patch set (#2). Change subject: Release notes for 0.9.1 .. Release notes for 0.9.1 Change-Id: I797980d105e0e1000420aedf58e094f9505053c6 --- M docs/installation.adoc M docs/release_notes.adoc 2 files changed, 34 insertions(+), 3 deletions(-) git pull ssh://gerrit.cloudera.org:29418/kudu refs/changes/63/3463/2 -- To view, visit http://gerrit.cloudera.org:8080/3463 To unsubscribe, visit http://gerrit.cloudera.org:8080/settings Gerrit-MessageType: newpatchset Gerrit-Change-Id: I797980d105e0e1000420aedf58e094f9505053c6 Gerrit-PatchSet: 2 Gerrit-Project: kudu Gerrit-Branch: branch-0.9.x Gerrit-Owner: Todd Lipcon Gerrit-Reviewer: Jean-Daniel Cryans Gerrit-Reviewer: Kudu Jenkins Gerrit-Reviewer: Misty Stanley-Jones Gerrit-Reviewer: Todd Lipcon
[kudu-CR](branch-0.9.x) Release notes for 0.9.1
Kudu Jenkins has posted comments on this change. Change subject: Release notes for 0.9.1 .. Patch Set 2: Build Started http://104.196.14.100/job/kudu-gerrit/1965/ -- To view, visit http://gerrit.cloudera.org:8080/3463 To unsubscribe, visit http://gerrit.cloudera.org:8080/settings Gerrit-MessageType: comment Gerrit-Change-Id: I797980d105e0e1000420aedf58e094f9505053c6 Gerrit-PatchSet: 2 Gerrit-Project: kudu Gerrit-Branch: branch-0.9.x Gerrit-Owner: Todd Lipcon Gerrit-Reviewer: Jean-Daniel Cryans Gerrit-Reviewer: Kudu Jenkins Gerrit-Reviewer: Misty Stanley-Jones Gerrit-Reviewer: Todd Lipcon Gerrit-HasComments: No
[kudu-CR](branch-0.9.x) KUDU-1473: fix some tablet lock usage in CatalogManager
Adar Dembo has submitted this change and it was merged. Change subject: KUDU-1473: fix some tablet lock usage in CatalogManager .. KUDU-1473: fix some tablet lock usage in CatalogManager This was probably due to the refactoring done in commit 59ff89d. Now that we're releasing tablet write locks as early as possible, we need to reacquire them (in read mode) for some operations. The lock misuse reared its head most often in the Java test TestKuduTable.testGetLocations, but could trigger in just about any test that didn't wait for table creation to finish before accessing the table. For example, CreateTableITest.TestCreateWhenMajorityOfReplicasFailCreation was a little flaky too. The backtrace was always the same: cow_object.h:82] Check failed: lock_.HasReaders() || lock_.HasWriteLock() @ 0x7fecb643e37b kudu::CowObject<>::state() at ??:0 @ 0x7fecb6422fe9 kudu::master::CatalogManager::ProcessPendingAssignments() at ??:0 @ 0x7fecb6421d3a kudu::master::CatalogManagerBgTasks::Run() at ??:0 @ 0x7fecb645c1f7 boost::_mfi::mf0<>::operator()() at ??:0 @ 0x7fecb645c15b boost::_bi::list1<>::operator()<>() at ??:0 @ 0x7fecb645c104 boost::_bi::bind_t<>::operator()() at ??:0 @ 0x7fecb645bf2a boost::detail::function::void_function_obj_invoker0<>::invoke() at ??:0 @ 0x7fecb0b66d82 boost::function0<>::operator()() at ??:0 @ 0x7fecafba96c0 kudu::Thread::SuperviseThread() at ??:0 @ 0x423faa __tsan_thread_start_func at ??:0 @ 0x7fecb28459d1 start_thread at ??:0 @ 0x7fecacd108fd clone at ??:0 Note: on branch-0.9.x, SendDeleteTabletRequest() does not need to acquire a tablet lock because it still uses the per-tablet replica cache which has since been removed on master. Change-Id: I8e24f6035f4d778995ea3f295396f5fbd760d6c6 Reviewed-on: http://gerrit.cloudera.org:8080/3309 Reviewed-by: Mike Percy Tested-by: Adar Dembo (cherry picked from commit 679c6b2c93313bd4371916678b4ea8f52b621a50) Reviewed-on: http://gerrit.cloudera.org:8080/3460 Tested-by: Kudu Jenkins --- M src/kudu/master/catalog_manager.cc M src/kudu/master/catalog_manager.h 2 files changed, 33 insertions(+), 22 deletions(-) Approvals: Mike Percy: Looks good to me, approved Kudu Jenkins: Verified -- To view, visit http://gerrit.cloudera.org:8080/3460 To unsubscribe, visit http://gerrit.cloudera.org:8080/settings Gerrit-MessageType: merged Gerrit-Change-Id: I8e24f6035f4d778995ea3f295396f5fbd760d6c6 Gerrit-PatchSet: 3 Gerrit-Project: kudu Gerrit-Branch: branch-0.9.x Gerrit-Owner: Adar Dembo Gerrit-Reviewer: Adar Dembo Gerrit-Reviewer: Kudu Jenkins Gerrit-Reviewer: Mike Percy
[kudu-CR] Allow for reserving disk space for non-Kudu processes
Kudu Jenkins has posted comments on this change. Change subject: Allow for reserving disk space for non-Kudu processes .. Patch Set 8: Build Started http://104.196.14.100/job/kudu-gerrit/1964/ -- To view, visit http://gerrit.cloudera.org:8080/3135 To unsubscribe, visit http://gerrit.cloudera.org:8080/settings Gerrit-MessageType: comment Gerrit-Change-Id: Ifd0451d4dbddc1783019a53302de0263080939c7 Gerrit-PatchSet: 8 Gerrit-Project: kudu Gerrit-Branch: master Gerrit-Owner: Mike Percy Gerrit-Reviewer: Adar Dembo Gerrit-Reviewer: Kudu Jenkins Gerrit-Reviewer: Mike Percy Gerrit-Reviewer: Todd Lipcon Gerrit-HasComments: No
[kudu-CR](branch-0.9.x) KUDU-1477. Pending COMMIT message for failed write operation can prevent tablet startup
Todd Lipcon has posted comments on this change. Change subject: KUDU-1477. Pending COMMIT message for failed write operation can prevent tablet startup .. Patch Set 2: Code-Review+2 -- To view, visit http://gerrit.cloudera.org:8080/3464 To unsubscribe, visit http://gerrit.cloudera.org:8080/settings Gerrit-MessageType: comment Gerrit-Change-Id: I8ecf8d780de1aa89fae4e0510d8291eb1f1cee11 Gerrit-PatchSet: 2 Gerrit-Project: kudu Gerrit-Branch: branch-0.9.x Gerrit-Owner: Todd Lipcon Gerrit-Reviewer: David Ribeiro Alves Gerrit-Reviewer: Kudu Jenkins Gerrit-Reviewer: Todd Lipcon Gerrit-HasComments: No
[kudu-CR](branch-0.9.x) KUDU-1477. Pending COMMIT message for failed write operation can prevent tablet startup
Todd Lipcon has submitted this change and it was merged. Change subject: KUDU-1477. Pending COMMIT message for failed write operation can prevent tablet startup .. KUDU-1477. Pending COMMIT message for failed write operation can prevent tablet startup This fixes a bug seen in a recent YCSB stress test that I ran in which I was accidentally writing tens of thousands of duplicate keys per second. After a tablet server restarted, it failed to come up due to a pending commit which referred to no mutated stores (e.g. because all of the operations were duplicate key inserts). This patch tweaks the logic for this safety check: a commit with no mutated stores trivially has "no active stores". However, that's not the same as having "only inactive stores" -- the subtlety is in the difference in behavior when a commit has no stores at all. The patch adds a new targeted test in tablet_bootstrap-test as well as a more end-to-end test in ts_recovery-itest. Both reproduced the bug reliably before this patch. Change-Id: I8ecf8d780de1aa89fae4e0510d8291eb1f1cee11 Reviewed-on: http://gerrit.cloudera.org:8080/3321 Tested-by: Kudu Jenkins Reviewed-by: David Ribeiro Alves (cherry picked from commit 6894438a406a635dc8a8f3bd77862294163cc7fb) Reviewed-on: http://gerrit.cloudera.org:8080/3464 Reviewed-by: Todd Lipcon --- M src/kudu/consensus/log-test-base.h M src/kudu/integration-tests/raft_consensus-itest.cc M src/kudu/integration-tests/test_workload.cc M src/kudu/integration-tests/test_workload.h M src/kudu/integration-tests/ts_recovery-itest.cc M src/kudu/tablet/tablet_bootstrap-test.cc M src/kudu/tablet/tablet_bootstrap.cc 7 files changed, 148 insertions(+), 23 deletions(-) Approvals: Todd Lipcon: Looks good to me, approved Kudu Jenkins: Verified -- To view, visit http://gerrit.cloudera.org:8080/3464 To unsubscribe, visit http://gerrit.cloudera.org:8080/settings Gerrit-MessageType: merged Gerrit-Change-Id: I8ecf8d780de1aa89fae4e0510d8291eb1f1cee11 Gerrit-PatchSet: 3 Gerrit-Project: kudu Gerrit-Branch: branch-0.9.x Gerrit-Owner: Todd Lipcon Gerrit-Reviewer: David Ribeiro Alves Gerrit-Reviewer: Kudu Jenkins Gerrit-Reviewer: Todd Lipcon
[kudu-CR](branch-0.9.x) KUDU-1477. Pending COMMIT message for failed write operation can prevent tablet startup
Hello David Ribeiro Alves, Kudu Jenkins, I'd like you to reexamine a change. Please visit http://gerrit.cloudera.org:8080/3464 to look at the new patch set (#2). Change subject: KUDU-1477. Pending COMMIT message for failed write operation can prevent tablet startup .. KUDU-1477. Pending COMMIT message for failed write operation can prevent tablet startup This fixes a bug seen in a recent YCSB stress test that I ran in which I was accidentally writing tens of thousands of duplicate keys per second. After a tablet server restarted, it failed to come up due to a pending commit which referred to no mutated stores (e.g. because all of the operations were duplicate key inserts). This patch tweaks the logic for this safety check: a commit with no mutated stores trivially has "no active stores". However, that's not the same as having "only inactive stores" -- the subtlety is in the difference in behavior when a commit has no stores at all. The patch adds a new targeted test in tablet_bootstrap-test as well as a more end-to-end test in ts_recovery-itest. Both reproduced the bug reliably before this patch. Change-Id: I8ecf8d780de1aa89fae4e0510d8291eb1f1cee11 Reviewed-on: http://gerrit.cloudera.org:8080/3321 Tested-by: Kudu Jenkins Reviewed-by: David Ribeiro Alves (cherry picked from commit 6894438a406a635dc8a8f3bd77862294163cc7fb) --- M src/kudu/consensus/log-test-base.h M src/kudu/integration-tests/raft_consensus-itest.cc M src/kudu/integration-tests/test_workload.cc M src/kudu/integration-tests/test_workload.h M src/kudu/integration-tests/ts_recovery-itest.cc M src/kudu/tablet/tablet_bootstrap-test.cc M src/kudu/tablet/tablet_bootstrap.cc 7 files changed, 148 insertions(+), 23 deletions(-) git pull ssh://gerrit.cloudera.org:29418/kudu refs/changes/64/3464/2 -- To view, visit http://gerrit.cloudera.org:8080/3464 To unsubscribe, visit http://gerrit.cloudera.org:8080/settings Gerrit-MessageType: newpatchset Gerrit-Change-Id: I8ecf8d780de1aa89fae4e0510d8291eb1f1cee11 Gerrit-PatchSet: 2 Gerrit-Project: kudu Gerrit-Branch: branch-0.9.x Gerrit-Owner: Todd Lipcon Gerrit-Reviewer: David Ribeiro Alves Gerrit-Reviewer: Kudu Jenkins
[kudu-CR](branch-0.9.x) KUDU-1477. Pending COMMIT message for failed write operation can prevent tablet startup
Kudu Jenkins has posted comments on this change. Change subject: KUDU-1477. Pending COMMIT message for failed write operation can prevent tablet startup .. Patch Set 2: Build Started http://104.196.14.100/job/kudu-gerrit/1963/ -- To view, visit http://gerrit.cloudera.org:8080/3464 To unsubscribe, visit http://gerrit.cloudera.org:8080/settings Gerrit-MessageType: comment Gerrit-Change-Id: I8ecf8d780de1aa89fae4e0510d8291eb1f1cee11 Gerrit-PatchSet: 2 Gerrit-Project: kudu Gerrit-Branch: branch-0.9.x Gerrit-Owner: Todd Lipcon Gerrit-Reviewer: David Ribeiro Alves Gerrit-Reviewer: Kudu Jenkins Gerrit-HasComments: No