Created
May 9, 2018 22:21
-
-
Save steveyen/6f64028a23326b6c4c554a2c539787ad to your computer and use it in GitHub Desktop.
getchangelog?product=couchbase-server&fromb=5.5.0-2630&tob=5.5.0-2671
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| BuildBoard | |
| Build Status | |
| Change Log | |
| CHANGELOG for asterix-opt | |
| * Commit: dd30f909eed03d2b6d4e316850477171ff487786 | |
| MB-26884, MB-2658, MB-26743, and docs for convergence | |
| Change-Id: If0f18bd1d59e5380fd537f358e9e4bfa909fb736 | |
| Reviewed-on: http://review.couchbase.org/93633 | |
| Reviewed-by: Murtadha Hubail <[email protected]> | |
| Tested-by: Murtadha Hubail <[email protected]> | |
| * Commit: 326322b03af4912bcd4d0c8bb1369b21e8971ea1 | |
| MB-29503: use repo manifest -r for version information | |
| The solution (repo forall-based) to determine version information | |
| from repo state is not working on the version of repo available on | |
| Windows. Rework the solution to instead use repo manifest -r, | |
| which is working on all known platforms. | |
| Change-Id: I205377da04c840f0d2885e2cb065e48f39dcd29d | |
| Reviewed-on: http://review.couchbase.org/93707 | |
| Reviewed-by: Murtadha Hubail <[email protected]> | |
| Tested-by: Michael Blow <[email protected]> | |
| * Commit: 648c8136515f4c36b853d6a56d8532d46dc241cf | |
| Revert "MB-29503: use repo manifest -r for version information" | |
| This reverts commit 7ae8c967a1447f999230187a24e1821fcc0ed2ec. | |
| Change-Id: I92fb363ab2f40e78a386032e4e0ab976b704f995 | |
| Reviewed-on: http://review.couchbase.org/93705 | |
| Reviewed-by: Michael Blow <[email protected]> | |
| Tested-by: Michael Blow <[email protected]> | |
| * Commit: 7ae8c967a1447f999230187a24e1821fcc0ed2ec | |
| MB-29503: use repo manifest -r for version information | |
| The solution (repo forall-based) to determine version information | |
| from repo state is not working on the version of repo available on | |
| Windows. Rework the solution to instead use repo manifest -r, | |
| which is working on all known platforms. | |
| Change-Id: Idbeaf181c07b84f035990a7dbba14f9965f362d4 | |
| Reviewed-on: http://review.couchbase.org/93635 | |
| Reviewed-by: Murtadha Hubail <[email protected]> | |
| Tested-by: Michael Blow <[email protected]> | |
| * Commit: 1bb1f467887c86d338a064a8abc7a960e6f48a37 | |
| Move Analytics Dirs Outside Data Dir | |
| Change-Id: If9e45b3f1232bff56348e4e5542ce5f61a6714da | |
| Reviewed-on: http://review.couchbase.org/93665 | |
| Tested-by: Build Bot <[email protected]> | |
| Reviewed-by: Murtadha Hubail <[email protected]> | |
| Reviewed-by: Abdullah Amoudi <[email protected]> | |
| * Commit: 13dd912b957c7a1e9e32338d90e74cdbbb974da0 | |
| Correct using downloaded JDK | |
| Maven 3.5.2 does not honor JAVACMD in the environment, but instead wants | |
| JAVA_HOME set appropriately. Switch to discovering JAVA_HOME and passing | |
| it to Maven to ensure that it uses the JDK we download. | |
| Also remove unused code in FindCouchbaseMaven.cmake to avoid confusion, | |
| and fix typo causing a CMake warning. | |
| Change-Id: I96319b43a73344b226a4361c69a499d5d84aa81d | |
| Reviewed-on: http://review.couchbase.org/93598 | |
| Tested-by: Build Bot <[email protected]> | |
| Reviewed-by: Michael Blow <[email protected]> | |
| Reviewed-by: Kenneth Lareau <[email protected]> | |
| CHANGELOG for backup | |
| * Commit: 780a126f15ccd4f1f71d1ad120522b0dbc98803f | |
| MB-29323: Fix broken unit tests | |
| Change-Id: I1f163a8236a4a8a1466855f5e475a3fef990caae | |
| Reviewed-on: http://review.couchbase.org/93664 | |
| Tested-by: Mike Wiederhold <[email protected]> | |
| Reviewed-by: Patrick Varley <[email protected]> | |
| * Commit: 98666349e806c422bf6c234e4b5563ed5e7bd759 | |
| MB-29440 cbimport print a summary | |
| cbimport prints the number of successful and unsuccessful | |
| documents imported into the cluster. | |
| Change-Id: I61b1674c186332d5b1d7d80b2bba1d9f86547071 | |
| Reviewed-on: http://review.couchbase.org/93426 | |
| Tested-by: Patrick Varley <[email protected]> | |
| Reviewed-by: Mike Wiederhold <[email protected]> | |
| CHANGELOG for bleve | |
| * Commit: ecf672f9bf46edfafa0262cbe05cc943b72ff48b | |
| Merge pull request #926 from mschoch/fix-fuzzy-and-related | |
| Fix fuzzy and related | |
| * Commit: b1b570b3d9b0e359833ea7f0819bd7ab87a08490 | |
| return err and close properly for fuzzy and regexp search | |
| * Commit: e52660b99369b0e79f3de17434f5c7594aa61986 | |
| correctly return dictionary iterator creation errors in Next() | |
| * Commit: ebdbbffd06d86d411f1ba4fbbc10c94370e6ddb4 | |
| limit fuzzy edit distance to 2 | |
| * Commit: f83cd37e6a80e2d06a38418bef3b97af066b42f7 | |
| Merge pull request #916 from sreekanth-cb/dvreader_version_1 | |
| dvReader improvement | |
| * Commit: a8bd3e641e1ead048ded267094922f69b68ca44b | |
| refactoring the dvReader reuse | |
| * Commit: b22d4fd4eff3e850a9e508c84e0e72300636bcab | |
| Merge pull request #922 from abhinavdangeti/master | |
| Micro optimizations while estimating mem needed for search | |
| * Commit: 7c996711e7dadf65200295244305ba3a695be1c5 | |
| Micro optimizations while estimating mem needed for search | |
| * Commit: 612a79136d2767e907a6724aa9fe582efe4f6d18 | |
| Merge pull request #921 from steveyen/optimize-scorch-ExternalID | |
| scorch optimize ExternalID() with Segment.DocID() API | |
| * Commit: d784dda7525192a4f8d528e75798760529ee950c | |
| Merge pull request #920 from steveyen/optimize-via-vellum-FTS-Reader-API | |
| scorch optimize via vellum.FST.Reader() API | |
| * Commit: 23aaeb730497c01ec0dfab0f05cc6adb62097b1e | |
| scorch optimize ExternalID() with Segment.DocID() API | |
| The previous implementation of ExternalID(), which used the generic | |
| VisitDocument() API to access the _id field (with extra closure | |
| allocations), has now been replaced in this optimization by invoking a | |
| new Segment.DocID() method. | |
| * Commit: 0a3f3e44224c1fc8f8b0e8777b7554abb8113d3b | |
| scorch optimize via vellum.FST.Reader() API | |
| * Commit: 96dad7fda05b25b250b4ffe0e593d4e293b96e21 | |
| dvReader improvement | |
| -attempt to improve the reuse of dvReaders | |
| -avoid the redundant dvCache prep checks | |
| CHANGELOG for cbft | |
| * Commit: 54a9d48c44108bb4c7bd79ee335a3ff301729d87 | |
| MB-29504 - short term fix to prevent total indexing blockage | |
| Recent changes to memory accounting allowed for the possibility | |
| that the app_herder would block indexing forever. This change | |
| is an attempt to allow indexing to slowly continue in these | |
| cases. | |
| Change-Id: Idba5b72273de74efd555c5c3591c9db5c8ed3c9a | |
| Reviewed-on: http://review.couchbase.org/93776 | |
| Well-Formed: Build Bot <[email protected]> | |
| Reviewed-by: Abhinav Dangeti <[email protected]> | |
| Tested-by: Marty Schoch <[email protected]> | |
| * Commit: 3233dc3c35bd53a9799e9d60a2367473deb5c62a | |
| MB-29142: Add IdleConnTimeout for http2 client's transport | |
| Change-Id: I81358ce1b5152fd1a1b352525225c9c9d431ea81 | |
| Reviewed-on: http://review.couchbase.org/93638 | |
| Well-Formed: Build Bot <[email protected]> | |
| Reviewed-by: Steve Yen <[email protected]> | |
| Tested-by: Abhinav Dangeti <[email protected]> | |
| * Commit: b66521fd5b89b82d7a93224a038527e619133137 | |
| MB-29142: Use HTTP transport, and configure to HTTP/2 | |
| Change-Id: I136b943d5b3f229ab8a113609bf26e7b45cac876 | |
| Reviewed-on: http://review.couchbase.org/93592 | |
| Well-Formed: Build Bot <[email protected]> | |
| Reviewed-by: Steve Yen <[email protected]> | |
| Tested-by: Abhinav Dangeti <[email protected]> | |
| * Commit: d4acfdbee31b1be8a2f721b89545b98811c8fb28 | |
| Plugging in scorch's rollback routines into FTS | |
| As part of this change moved out the common stuff | |
| from the current rollback procedure, and created | |
| separate files for upsidedown/moss's rollback and | |
| scorch's rollback. | |
| Change-Id: If13a46fb34a5cd8ef42bc2b4ffbd0a3dd0afc1bb | |
| Reviewed-on: http://review.couchbase.org/88004 | |
| Reviewed-by: Marty Schoch <[email protected]> | |
| Well-Formed: Build Bot <[email protected]> | |
| Reviewed-by: Steve Yen <[email protected]> | |
| Tested-by: Abhinav Dangeti <[email protected]> | |
| * Commit: 60742b77a19642950187d020ecfc080aad1cc635 | |
| MB-29491: Upgrade Go version to 1.9.6 | |
| Addresses string comparison issue: | |
| https://github.com/golang/go/issues/24817 | |
| Change-Id: I8a8a10f2739e8b1a4884f3b74db7a6b0db78f35b | |
| Reviewed-on: http://review.couchbase.org/93589 | |
| Well-Formed: Build Bot <[email protected]> | |
| Reviewed-by: Marty Schoch <[email protected]> | |
| Tested-by: Abhinav Dangeti <[email protected]> | |
| CHANGELOG for couchbase-cli | |
| * Commit: 7142291d009034f4ed13bcad0fa0425ac3bd3d16 | |
| MB-27925 Fix CE topology enforcement | |
| Change-Id: Ia78bc1c1f1b65ac7de83de245a5cd116b67942c0 | |
| Reviewed-on: http://review.couchbase.org/93697 | |
| Reviewed-by: Mike Wiederhold <[email protected]> | |
| Tested-by: Patrick Varley <[email protected]> | |
| * Commit: a7bfd9863610811facbdf70523fbc94431ddde89 | |
| MB-29432: Properly set REST parameters for enable-email-encrypt | |
| Change-Id: Ic2ad6f0d5135417f5a8b2803d02d55d6adad1bd5 | |
| Reviewed-on: http://review.couchbase.org/93646 | |
| Reviewed-by: Patrick Varley <[email protected]> | |
| Tested-by: Patrick Varley <[email protected]> | |
| CHANGELOG for couchstore | |
| * Commit: 28ec7f5b4e2dca0e7496396f60a657edb1abf9d8 | |
| Avoid undefined behaviour for zero-length rev-meta | |
| As identified by UBSan when running couchstore-purge test: | |
| runtime error: null pointer passed as argument 2, which is declared to never be null | |
| #0 0x7f4c2e622f88 in assemble_id_index_value couchstore/src/couch_save.cc:57 | |
| #1 0x7f4c2e622f88 in add_doc_to_update_list couchstore/src/couch_save.cc:326 | |
| #2 0x7f4c2e622f88 in couchstore_save_documents couchstore/src/couch_save.cc:387 | |
| Change-Id: Ic759ba8d455ddc048636269ad0c051808164407d | |
| Reviewed-on: http://review.couchbase.org/93770 | |
| Tested-by: Build Bot <[email protected]> | |
| Reviewed-by: Trond Norbye <[email protected]> | |
| * Commit: e0680f06905bc0f3aca3c0b70ce9db019ce68b3f | |
| Avoid undefined behaviour for zero-length reduce value | |
| As reported by UBSan, it is undefined behviour to call memcpy() with a | |
| null argument - in this case when encoding a root B-Tree node which | |
| has no reduce value (local tree has an empty reduce value). | |
| Change-Id: I5fc62025b60039b799dbb78b30f4e406c1044951 | |
| Reviewed-on: http://review.couchbase.org/93750 | |
| Tested-by: Build Bot <[email protected]> | |
| Reviewed-by: Trond Norbye <[email protected]> | |
| * Commit: 8de65b7e13cd1e74e89457fbeceadf538ccc29a7 | |
| Avoid undefined behaviour for zero-length Document | |
| As reported by UBSan, it is undefined behaviour to call memcpy() with | |
| a null argument - in this case when Document has zero-length data. | |
| Change-Id: Iabba97b79185eb4d5d4c1d8e73a421fbc1e19e47 | |
| Reviewed-on: http://review.couchbase.org/93749 | |
| Tested-by: Build Bot <[email protected]> | |
| Reviewed-by: Trond Norbye <[email protected]> | |
| * Commit: 16455239689ce664fb692f861632de2720e9c557 | |
| dbdiff: Correctly compare compressed values | |
| Compressed values were previously not decompressed before comparing | |
| their length or content. | |
| This resulted in spurious differences being reported; as when | |
| documents are Snappy-compressed there's no guarantee that the exact | |
| same compressed encoding will always be used. This manifested in the | |
| size of compressed values sometimes differing by one. | |
| Fix this by comparing the uncompressed length. | |
| Additionally now we have the uncompressed value we can correctly | |
| compare the (uncompressed) values. | |
| Change-Id: I2b1a548091db728ce4d10bf6f874bbf6bd6c060b | |
| Reviewed-on: http://review.couchbase.org/93545 | |
| Reviewed-by: Trond Norbye <[email protected]> | |
| Tested-by: Build Bot <[email protected]> | |
| * Commit: 93ee71d9fa8f395fd0ca1a506129c97423f8c653 | |
| dbdiff: compare couchbase per-revision metdata | |
| Expand dbdiff to compare the couchstore metadata (CAS, expiry, flags) | |
| of each document. | |
| Change-Id: If99476ce4354bb4820a5929b7ba3053ee5ff8363 | |
| Reviewed-on: http://review.couchbase.org/93544 | |
| Reviewed-by: Trond Norbye <[email protected]> | |
| Tested-by: Build Bot <[email protected]> | |
| * Commit: 1370ab0052b0a0177864fea21ceec6b7326c2b2f | |
| dbdiff: Don't ignore differences in quiet mode | |
| When dbdiff is run in quiet mode, it incorreclty skipped differences | |
| when comparing docinfo objects. Quiet mode should only skip printing | |
| the details; it shouldn't skip actually performing the checks. | |
| Change-Id: Ic3ee7c5977977f78fff4e126a32e94015eb0f940 | |
| Reviewed-on: http://review.couchbase.org/93612 | |
| Reviewed-by: Trond Norbye <[email protected]> | |
| Tested-by: Build Bot <[email protected]> | |
| * Commit: 3e1078ad7d72422b4a3affd8d557d7f5670dc7e9 | |
| dbdiff: Output messages to stdout | |
| Change dbdiff so the details of the differences between couchstore | |
| files is output to stdout, instead of stderr. This makes it easier to | |
| extract / manipulate the output of the program (e.g. piping to grep). | |
| Change-Id: I0f88a657071c136191fdad3206e7d47f02ed37dc | |
| Reviewed-on: http://review.couchbase.org/93543 | |
| Tested-by: Build Bot <[email protected]> | |
| Reviewed-by: Jim Walker <[email protected]> | |
| Reviewed-by: Trond Norbye <[email protected]> | |
| * Commit: 3edac8daa9c819f5744612dd66fcffc8123cbc72 | |
| dbdiff: Convert to C++ | |
| Motivation for this is that I want to use C++ code (bitfields.h) in | |
| dbdiff. | |
| Change-Id: I6e5fb553e202127e0250f6fe1ade5e2dec94e708 | |
| Reviewed-on: http://review.couchbase.org/93542 | |
| Reviewed-by: Trond Norbye <[email protected]> | |
| Tested-by: Build Bot <[email protected]> | |
| * Commit: a2cae2dce997f7f81132fa4a57c065fb06fe221e | |
| dbdiff: Correctly report deleted docs | |
| dbdiff is incorrectly reporting that deleted documents are missing | |
| from second couchstore file; when the document is deleted in both | |
| files. | |
| This is because if a document has been deleted; then it cannot be | |
| fetched using couchstore_open_document(). Instead one must first call | |
| couchstore_docinfo_by_id() to fetch the docinfo; and then examine | |
| DocInfo.deleted. | |
| Change-Id: Iedb0e4c206806e5286d110dccba1eb2e24b25a40 | |
| Reviewed-on: http://review.couchbase.org/93538 | |
| Tested-by: Build Bot <[email protected]> | |
| Reviewed-by: Tim Bradgate <[email protected]> | |
| CHANGELOG for go-couchbase | |
| * Commit: 16db1f1fe037412f12738fa4d8448c549c4edd77 | |
| MB-29471, MB-29034 Use SetDeadline. Instead of Read/Write | |
| When SetReadDedline(), SetWriteDeadline() both are set separately the number of | |
| timers are double and causes lock contentions. If only one is set it is fine. | |
| Use SetDeadline() which is for both read and write | |
| Change-Id: If624876afa4bc18432f6ace450ae7d5398417124 | |
| Reviewed-on: http://review.couchbase.org/93523 | |
| Reviewed-by: Pavel Paulau <[email protected]> | |
| Reviewed-by: Marco Greco <[email protected]> | |
| Tested-by: Sitaram Vemulapalli <[email protected]> | |
| CHANGELOG for gomemcached | |
| * Commit: 0da75df145308b9a4e6704d762ca9d9b77752efc | |
| MB-100 - regenerate mock based on latest interface | |
| Change-Id: Ifc1be9245b219245cb7c6260a443ce9eeb21f30e | |
| * Commit: bf40830d1642a108b93fe3ae5bc2ae1a703ec9f9 | |
| Merge "MB-28759 - Add start marker for compression type" | |
| * Commit: a0f176b5fc3a4b915b4e7491ccdeefda7e75fc38 | |
| MB-29471, MB-29034 Use SetDeadline. Instead of Read/Write | |
| Change-Id: I9a936c83b0f320c8a5e190d6a6f76e3a65dcd811 | |
| * Commit: 72a8e5737cc10fd0d0a9f65396d8f5674ab32796 | |
| MB-28759 - Add start marker for compression type | |
| Change-Id: I45cbf4051e494ab7fdb09f10deadefe26281eaf7 | |
| CHANGELOG for goxdcr | |
| * Commit: ab511b8aec737e02150d13a2363f3c91be7b35bc | |
| MB-29477 - increase http server read timeout | |
| Change-Id: Id517323be12b5193781853e59749bcf405a2f70a | |
| Reviewed-on: http://review.couchbase.org/93627 | |
| Reviewed-by: Yu Sui <[email protected]> | |
| Reviewed-by: Timofey Barmin <[email protected]> | |
| Reviewed-by: Neil Huang <[email protected]> | |
| Tested-by: Neil Huang <[email protected]> | |
| * Commit: 786a66d2bbe16d646618f449cfef36eade4b21fe | |
| MB-29489 fix race condition in removeSpecFromCache | |
| Change-Id: I8020c616d0ff383afcef841fd9fe07b12738228a | |
| Reviewed-on: http://review.couchbase.org/93561 | |
| Reviewed-by: Neil Huang <[email protected]> | |
| Reviewed-by: Yu Sui <[email protected]> | |
| Tested-by: Yu Sui <[email protected]> | |
| * Commit: 113c5c21619bae965d4799d8d8528546fcabc25a | |
| MB-28759 - Change compression setting to default to Auto | |
| Change-Id: Ic4742fd12da269afbf89586dec53ca10be19b5ad | |
| Reviewed-on: http://review.couchbase.org/92736 | |
| Reviewed-by: Yu Sui <[email protected]> | |
| Reviewed-by: Neil Huang <[email protected]> | |
| Tested-by: Neil Huang <[email protected]> | |
| * Commit: 6b9a2ce0b728df89b8b93378c11c13a439bba6a1 | |
| MB-29477 - GoXDCR server should honor TCP keepalive | |
| Change-Id: I40fa01492f91ac96ce0ea1c398d9a7c3114d2d48 | |
| Reviewed-on: http://review.couchbase.org/93560 | |
| Reviewed-by: Yu Sui <[email protected]> | |
| Reviewed-by: Neil Huang <[email protected]> | |
| Tested-by: Neil Huang <[email protected]> | |
| * Commit: a214d1e729d5095241517b0c54d539b69a9757ed | |
| MB-29490 fix race condition on stats_interval | |
| Change-Id: Icb7a837707a9a97f816adcf5caf45f180e878436 | |
| Reviewed-on: http://review.couchbase.org/93580 | |
| Reviewed-by: Neil Huang <[email protected]> | |
| Reviewed-by: Yu Sui <[email protected]> | |
| Tested-by: Yu Sui <[email protected]> | |
| * Commit: 6992510875730923cf9dfc6759be5994bb3ecd13 | |
| MB-29488 fix race condition involving vb list | |
| Change-Id: Id2f8904bc7e7ad6ef735ca8239cc4fb9050a46df | |
| Reviewed-on: http://review.couchbase.org/93559 | |
| Reviewed-by: Neil Huang <[email protected]> | |
| Reviewed-by: Yu Sui <[email protected]> | |
| Tested-by: Yu Sui <[email protected]> | |
| CHANGELOG for indexing | |
| * Commit: 214e26e8bb7706454c58eec4d5895af41124c64c | |
| Merge commit '9fe5eda' into HEAD | |
| http://ci2i-unstable.northscale.in/gsi-fdb-04.05.2018-13.32.pass.html | |
| * Commit: 9fe5eda7f7f79e34eb0d3b5929be21d0e95a318f | |
| MB-29389: Fix hang for plasma rollback to 0 with concurrent partition index scan | |
| Change-Id: I51357f9870eabdd5e4d772edf2dee61240ea1886 | |
| * Commit: 1e3044f897880acb39b039ab6dccb42809937e6f | |
| Revert "MB-29317: plasma - serialize persistence by instance" | |
| This reverts commit 09d0be19d4cd887a587c8d401be11436cd975427. | |
| * Commit: eb577b206a7aba57ae1313964a6e783f6525d290 | |
| Revert "MB-100 picking up eventing unstable branch in GSI CI builds" | |
| This reverts commit 1682ca61db4547119ddaaed849c9f408a8b6e962. | |
| Change-Id: I643c415a7cc9a50b9f3dc0de79f3de1d766dfd72 | |
| * Commit: 1682ca61db4547119ddaaed849c9f408a8b6e962 | |
| MB-100 picking up eventing unstable branch in GSI CI builds | |
| Change-Id: Ifd111f2e9472924034bbe39e1f2a13988c948e42 | |
| * Commit: fbb2b69618b8fd602bde3adf4d874ea06730981b | |
| Merge commit 'afb804b' into HEAD | |
| http://ci2i-unstable.northscale.in/gsi-memdb-02.05.2018-01.04.pass.html | |
| * Commit: 09d0be19d4cd887a587c8d401be11436cd975427 | |
| MB-29317: plasma - serialize persistence by instance | |
| When a large number of indexes start writing to disk simultaneously, | |
| it causes disk to be 100% utilized which impacts query latency when | |
| node is in DGM. | |
| First control concurrency of persistence by persistenceCPUPercent | |
| Now, since persistence is already parallelized, | |
| only allow 1 index to perform full persistence at a given time. | |
| Change-Id: Icb8f7d0f34b0238b464fa37e1856d429f7a002c6 | |
| * Commit: afb804b0b9fb5841c9f2dc8e9d48ce7a897c816e | |
| MB-29368: Support discretionary compaction | |
| Change-Id: Ia4f3bcdf9711f82b010a2853b10ff2729029ad2b | |
| * Commit: 9b71a21da0d07a96e98f554f4599bd5de5b0f38d | |
| MB-29368: Enable Plasma manual compaction | |
| Change-Id: I5ac6b1ca2e0b8a44405ce504a57af5f28c76b98e | |
| * Commit: 50337224087258d2882618e23338d2c81242613f | |
| MB-29368: Add partitionId to MsgIndexCompact | |
| Change-Id: I29a9251f2425bc66657ddf1cc4fc0bff87af5c0f | |
| * Commit: 246bf8641733fd245972785ee43d83b78065f057 | |
| MB-29368: Pass stats to compaction manager | |
| Change-Id: Ie67c6be9abb83e7126228ffdaa47b51c79ced899 | |
| * Commit: 1a8e509012b612efaf8f902ef8bf8c1e13247593 | |
| MB-29368: Refactor fdb compaction | |
| Change-Id: I1c37286e08c11487d3ce689123a3f827711824c1 | |
| * Commit: d727e5d211eae9f20f2ac5b240cc0f9e02cfe20a | |
| MB-29476 Increment projector version | |
| Change-Id: I9ecc6339357be153b7d4f7f0289375324a0f17ce | |
| * Commit: dcfe3f93ab462293f8577986bb4e7cdeac113453 | |
| MB-29132: Initialize plasma slice writer based on num partitions | |
| Change-Id: Id23b1a5a5c7333cfdde2fddfcb6a5b8e4843f5a6 | |
| CHANGELOG for kv_engine | |
| * Commit: 2a69644fe890ed689ff315303bd16c506a7aff7f | |
| memcached_testapp: Increase gtest_discover_tests timeout to 30s | |
| Increase the time gtest_discover_tests() allows for a test executable | |
| to run to discover the set of tests from the default (5s) to 30s, as | |
| we've seen this hit the timeout in CV. | |
| Change-Id: I8169fc8fed9b2414c1766e211ed4ad70dfba2a3d | |
| Reviewed-on: http://review.couchbase.org/93767 | |
| Tested-by: Build Bot <[email protected]> | |
| Reviewed-by: Trond Norbye <[email protected]> | |
| * Commit: 42f2369732297ec1341efd7bacf8de1d3c6e8f8e | |
| MB-29527: subdoc: Avoid undefined behaviour in operate_single_doc() | |
| As identified by UBSan, if a sub-document operation results in a | |
| zero-length result (which is valid); the current implementation passes | |
| a null pointer to memcpy, which is undefined behaviour: | |
| [ RUN ] TransportProtocols/XattrTest.SetXattrAndDeleteBasic/Mcbp_XattrYes_JsonYes_SnappyYes | |
| runtime error: null pointer passed as argument 2, which is declared to never be null | |
| #0 0xd32951 in operate_single_doc kv_engine/daemon/subdocument.cc:776 | |
| #1 0xd3522d in do_body_phase kv_engine/daemon/subdocument.cc:1136 | |
| #2 0xd3522d in subdoc_operate kv_engine/daemon/subdocument.cc:1183 | |
| #3 0xd3522d in subdoc_executor kv_engine/daemon/subdocument.cc:431 | |
| Fix by using std::copy instead. | |
| Change-Id: Ia5e4d7f76fd57a81c62b930ded7b85dd31a1ae24 | |
| Reviewed-on: http://review.couchbase.org/93766 | |
| Tested-by: Build Bot <[email protected]> | |
| Reviewed-by: Trond Norbye <[email protected]> | |
| * Commit: 7343e0d8127bbed983b5513156d6b009f4bf80ef | |
| ep_test_apis.cc: Avoid undefined behaviour upon zero length set | |
| It's undefined to pass a null pointer to memcpy(). Replace with std::copy. | |
| Change-Id: I5143bc0852f14d42ccf3f6199428f64e912ad3e1 | |
| Reviewed-on: http://review.couchbase.org/93765 | |
| Tested-by: Build Bot <[email protected]> | |
| Reviewed-by: Trond Norbye <[email protected]> | |
| * Commit: c0c1db19210aa948c97c843ca640de91baa60d66 | |
| kvstore_test: Set delete item to deleted | |
| In KVStoreParamTest/TestPersistenceCallbacksForDel; the item which was | |
| deleted was incorrectly not set to deleted. This is incorrect; the | |
| deleted flag should be true for any item which is a delete. | |
| Change-Id: Ibd63743de79f8f84a808638d0594e2614bef98af | |
| Reviewed-on: http://review.couchbase.org/93763 | |
| Tested-by: Build Bot <[email protected]> | |
| Reviewed-by: Trond Norbye <[email protected]> | |
| * Commit: acc67a3ab484988ce5d80007b5bc4ce5a457f832 | |
| MB-29522: CouchKVStore::rollback: check return result of readVBState | |
| As identified by UBSan, CouchKVStore::rollback() doesn't check the | |
| return value of resdVBState; which can result in it attempting to | |
| write a null local document to the rolled-back vBucket: | |
| [ RUN ] CouchKVStoreErrorInjectionTest.readVBState_open_local_document | |
| runtime error: null pointer passed as argument 2, which is declared to never be null | |
| #0 0x7ffff7b5f30a in encode_root couchstore/src/node_types.cc:75 | |
| #1 0x7ffff7b36033 in db_write_header couchstore/src/couch_db.cc:175 | |
| #2 0x7ffff7b3f487 in couchstore_commit couchstore/src/couch_db.cc:255 | |
| #3 0x12c0e6d in CouchKVStore::rollback(unsigned short, unsigned long, std::shared_ptr<RollbackCB>) kv_engine/engines/ep/src/couch-kvstore/couch-kvstore.cc:2674 | |
| #4 0xd15cc6 in CouchKVStoreErrorInjectionTest_readVBState_open_local_document_Test::TestBody() kv_engine/engines/ep/tests/module_tests/kvstore_test.cc:1030 | |
| The local document is essential to interpreting the vBucket file; so | |
| if we can't read it we need to fail the rollback. | |
| Change-Id: I83871de2d4a96197bce17cbc9f1147792795a783 | |
| Reviewed-on: http://review.couchbase.org/93761 | |
| Tested-by: Build Bot <[email protected]> | |
| Reviewed-by: Trond Norbye <[email protected]> | |
| * Commit: 38ba836040bb2aed2e64b7cfbc35b552e69ecac2 | |
| MB-29369: Obtain the streamMutex earlier in the snapshot task | |
| Obtain the streamMutex and also check the stream is in-memory / | |
| takeover-send before the task increments the cursor. | |
| Change-Id: I82ba9b959859921062f817f9f8e2c1452cb852e7 | |
| Reviewed-on: http://review.couchbase.org/93497 | |
| Reviewed-by: Dave Rigby <[email protected]> | |
| Tested-by: Build Bot <[email protected]> | |
| * Commit: a09cd0b6f96e1d3428f062d8bceb5e69edc16d41 | |
| MB-29515: Replace SpinLock with std::mutex | |
| This replaces SpinLock with std::mutex for all uses apart from | |
| Ephemeral's SequenceList[1]. | |
| Testing of this shows a ~5% improvement in throughput under daily perf. | |
| [1] Still need to test the effect of this change on Ephemeral, to see | |
| if it's a suitable change to make. | |
| Change-Id: I04d3b78bc34e2f880690b1b3e341decf8668ad09 | |
| Reviewed-on: http://review.couchbase.org/93694 | |
| Tested-by: Build Bot <[email protected]> | |
| Reviewed-by: Trond Norbye <[email protected]> | |
| * Commit: 2c0d03e1522a6980f6e1980b1332570f830984fa | |
| Merge branch 'spock' into vulcan | |
| * spock: | |
| MB-29483: Disable DCP cursor dropping | |
| MB-29287: Add dcp_reflection_test | |
| MB-29287: Move SynchronousEPEngine building to seperate method | |
| MB-29287: Add string / streaming methods for DCP responses | |
| Change-Id: I042e476069651df511dc9c0273ff053296dd2b1b | |
| * Commit: b8939ea2d04ab9c1a33334d19604c15a8ac1aad6 | |
| MB-29482: Merge branch '5.1.0' into spock | |
| * 5.1.0: | |
| MB-29483: Disable DCP cursor dropping | |
| MB-29287: Add dcp_reflection_test | |
| MB-29287: Move SynchronousEPEngine building to seperate method | |
| MB-29287: Add string / streaming methods for DCP responses | |
| Change-Id: Ia20000726e1d3959d9a8ee45c3f40489b72fb934 | |
| * Commit: 80665277773893b5e04ef817c4a4d6b719770c9e | |
| MB-29412 Use sendall instead of send | |
| The Python socket.send() will not send the whole payload in some | |
| situations, which will create a malformed packet. Using sendall() | |
| will ensure the whole payload is sent. | |
| Change-Id: Ib51b5b8ef33d764e600f73cf15bd9a1e9c155d22 | |
| Reviewed-on: http://review.couchbase.org/93608 | |
| Reviewed-by: Trond Norbye <[email protected]> | |
| Tested-by: Patrick Varley <[email protected]> | |
| Reviewed-by: Dave Rigby <[email protected]> | |
| * Commit: 11490babca4afaa225a072f44cd9e399ca95b790 | |
| MB-29483: Merge remote-tracking branch 'review/watson_ep' into 5.1.0 | |
| * review/watson_ep: | |
| MB-29483: Disable DCP cursor dropping | |
| MB-29287: Add dcp_reflection_test | |
| MB-29287: Move SynchronousEPEngine building to seperate method | |
| MB-29287: Add string / streaming methods for DCP responses | |
| Change-Id: I57cf5c902ac3eee7f4cf2eb25741fe9022679039 | |
| * Commit: 238af6b454e4fe2a08d9987fd9d572135848eb96 | |
| Merge remote-tracking branch 'ep-engine/watson' into kv_engine/watson_ep | |
| * ep-engine/watson: | |
| MB-29483: Disable DCP cursor dropping | |
| MB-29287: Add dcp_reflection_test | |
| MB-29287: Move SynchronousEPEngine building to seperate method | |
| MB-29287: Add string / streaming methods for DCP responses | |
| Change-Id: Ic4d36779ef5281cdc4cf2a3bc4275e291a5ad2c5 | |
| * Commit: aa9be740fe24bccd717a5539bbc9c3dffe978029 | |
| MB-28685: Don't run some tests when the compression mode is not 'off' | |
| Some tests explicitly require the compression mode to be 'off'. So, if | |
| the compression mode is not 'off', return without executing the test. | |
| Change-Id: I8974cbc71f2a78908ccbef1ec69bbd30f227a453 | |
| Reviewed-on: http://review.couchbase.org/93563 | |
| Reviewed-by: Jim Walker <[email protected]> | |
| Tested-by: Build Bot <[email protected]> | |
| * Commit: 09771ed1ab8aeda8ecb55f4dd461e477364bb15a | |
| Refactor: Wrap hdr_histogram in its own class | |
| Wraps the hdr_histogram data structure in its own class (HdrHistogram) | |
| thereby making it possible to use (for example to hold statistics) | |
| without needing to refer directly to the underlying C data structure. | |
| Change-Id: I4d4475cbbc41ff45250ba4940fe7b4e1cf2dd177 | |
| Reviewed-on: http://review.couchbase.org/93536 | |
| Reviewed-by: Dave Rigby <[email protected]> | |
| Tested-by: Build Bot <[email protected]> | |
| * Commit: 27780ed3d6dd8ea3154565662cf28c488ec16f8c | |
| MB-29040: [2/2] Sanitise delete with value via DcpConsumer | |
| 1) Allow a wider range of datatypes to be received by adjusting the | |
| mcbp validator. | |
| 2) Update the DcpConsumer so that delete with values are checked and | |
| sanitised, this means even a deleted marked as 'xattr' needs checking | |
| because the source may be sending an incorrect xattr with raw body + | |
| user xattrs which will need deleting. | |
| 3) Add a DcpConsumer test which checks we strip the faulty input. | |
| Change-Id: I219f21df9a63bc6b1c004fa382bd1f32f94a3e90 | |
| Reviewed-on: http://review.couchbase.org/93041 | |
| Reviewed-by: Dave Rigby <[email protected]> | |
| Tested-by: Build Bot <[email protected]> | |
| * Commit: 568615988f16b7a820b32895688aa60e3bda107b | |
| MB-29483: Disable DCP cursor dropping | |
| Disable DCP cursor dropping: | |
| 1. DCP replication consumers no longer request cursor dropping. | |
| 2. DCP producers default cursorDropping to false, and ignore any | |
| requests to enable it. | |
| Change-Id: Ia9b5620d43c821eeaa32a7ec67df419a51acc089 | |
| Reviewed-on: http://review.couchbase.org/93534 | |
| Reviewed-by: Tim Bradgate <[email protected]> | |
| Reviewed-by: Jim Walker <[email protected]> | |
| Well-Formed: Build Bot <[email protected]> | |
| Tested-by: Build Bot <[email protected]> | |
| * Commit: 0e543aa5e3f4e85e38cb7149fda78c26720802da | |
| Add VB high seqno to handle slow stream log message | |
| Change-Id: I0d7ba05ecf399303cbce5312c09c69bfaa503b6c | |
| Reviewed-on: http://review.couchbase.org/93489 | |
| Reviewed-by: Dave Rigby <[email protected]> | |
| Tested-by: Build Bot <[email protected]> | |
| * Commit: dfeff619171d43be0fd4e8277c15fa9fbaf01362 | |
| MB-29287: Add dcp_reflection_test | |
| Add new set of DCP unit tests, which connect DcpProducer and DcpConsumer | |
| objects; and reflect the messages between them. This allows us to test | |
| how the producer and consumer communicate; without involving | |
| ns_server. | |
| The tests are purely constrained to ep-engine - i.e. no packets are | |
| actually transmitted over the network; we instead just inject the | |
| messages directly into the DcpProducer / DcpConsumer objects. | |
| Change-Id: I641826238dc09c9b94d6540b8e4a5edc656883dc | |
| Reviewed-on: http://review.couchbase.org/92984 | |
| Well-Formed: Build Bot <[email protected]> | |
| Reviewed-by: Tim Bradgate <[email protected]> | |
| Reviewed-by: Jim Walker <[email protected]> | |
| Tested-by: Build Bot <[email protected]> | |
| * Commit: fab7de9d8e24ab904d1bbf1edabf06750143a81c | |
| MB-29287: Move SynchronousEPEngine building to seperate method | |
| Refactor the construction and setup of SynchronousEPEngine from | |
| EventuallyPersistentStoreTest::SetUp() to a seperate builder function | |
| - SynchronousEPEngine::build(). | |
| This allows tests which want to instantiate more than one ep-engine | |
| instance to do so. | |
| Change-Id: I647d32a28a2c3a1e1bfbca563c2a1ebe5130ad56 | |
| Reviewed-on: http://review.couchbase.org/93043 | |
| Well-Formed: Build Bot <[email protected]> | |
| Tested-by: Build Bot <[email protected]> | |
| Reviewed-by: Trond Norbye <[email protected]> | |
| * Commit: ca86e99d8c7765f908aaf285a3af566649c0577b | |
| MB-29287: Add string / streaming methods for DCP responses | |
| Allow printing of DCP response codes in log messages / test code. | |
| Change-Id: I7d48625fbf3efcad1cfb916686f585eec42e2b46 | |
| Reviewed-on: http://review.couchbase.org/92983 | |
| Well-Formed: Build Bot <[email protected]> | |
| Tested-by: Build Bot <[email protected]> | |
| Reviewed-by: Tim Bradgate <[email protected]> | |
| Reviewed-by: Trond Norbye <[email protected]> | |
| CHANGELOG for libcouchbase | |
| * Commit: 644132b44aa5bb9f93aca8bf9e3c2ebd2b190251 | |
| CCBC-925: Do not register nonexistent crypto providers during lookup | |
| As per previous commit, but using a copy of the iterator | |
| instead of a reference, in case of iterator lifetime issues. | |
| Change-Id: Id2da49fba8b76a7d1160fa7b6834b39f1787653c | |
| Reviewed-on: http://review.couchbase.org/93756 | |
| Reviewed-by: Sergey Avseyev <[email protected]> | |
| Tested-by: Build Bot <[email protected]> | |
| * Commit: 94bd4033d6555c5feecb98b5874c9b0ab65a7e05 | |
| CCBC-925: Do not register nonexistent crypto providers during lookup | |
| The std::map [] operator on the instance->crypto | |
| map was adding a NULL entry to the map when looking up | |
| provider names. | |
| This was leading to a crash on shutdown as the reference counts | |
| were being decremented on NULL entries. | |
| This fix checks for the existence of a provider using std::map::find | |
| rather than using the [] operator. | |
| Change-Id: I86f0b804b85fe4b9182de8f15fbf62f226e893e6 | |
| Reviewed-on: http://review.couchbase.org/93754 | |
| Tested-by: Build Bot <[email protected]> | |
| Reviewed-by: Sergey Avseyev <[email protected]> | |
| * Commit: b29cc33ccbcdd25f3e0d9b07d371e28a0267a2d3 | |
| CCBC-924: initialize flag for JSON server feature | |
| Otherwise it might be left uninitialized and the library will | |
| send JSON datatype to servers, which do not support it. | |
| Change-Id: Ib316816d1db4a7071a4864c7293f4ad374b01eac | |
| Reviewed-on: http://review.couchbase.org/93641 | |
| Tested-by: Build Bot <[email protected]> | |
| Reviewed-by: Pavel Paulau <[email protected]> | |
| Tested-by: Pavel Paulau <[email protected]> | |
| * Commit: 081e8b16b991bf706eb77f8243935c6fba31b895 | |
| Update release meta for 2.8.7 | |
| Change-Id: I50d65043f4fad4503d7441e600c4c7030224b20d | |
| Reviewed-on: http://review.couchbase.org/93605 | |
| Reviewed-by: Ellis Breen <[email protected]> | |
| Tested-by: Build Bot <[email protected]> | |
| * Commit: 4bbf137e778c31241f9170aa4b7ff7e1008e4b1d | |
| CCBC-923: Allow to disable fast-forward map for NMV handler | |
| Change-Id: I7beb8b87e914577e73e568b2938b1ec1f4814414 | |
| Reviewed-on: http://review.couchbase.org/93603 | |
| Tested-by: Build Bot <[email protected]> | |
| Reviewed-by: Sergey Avseyev <[email protected]> | |
| CHANGELOG for ns_server | |
| * Commit: 8b8ffa383b27b8f61ea1b052b8b0b25bced1fa0c | |
| Fix menelaus_web_node:validate_ix_cbas_path/2 | |
| returned wrong result for /abc/def and /abc/de | |
| Change-Id: Ib92b06fd1965dec36fcb1afb35d8c2f0e33b538b | |
| Reviewed-on: http://review.couchbase.org/93775 | |
| Well-Formed: Build Bot <[email protected]> | |
| Reviewed-by: Aliaksey Artamonau <[email protected]> | |
| Tested-by: Build Bot <[email protected]> | |
| * Commit: 5dc976c77acd44518dbd1f984ecab74fb65a8641 | |
| Merge remote-tracking branch 'couchbase/spock' into vulcan | |
| * couchbase/spock: | |
| MB-29436 Config upgrade to 5.1.1 | |
| MB-29436 Bump mb_master advertised version. | |
| MB-29436 Add a quirk to do trivial moves during rebalance. | |
| MB-29436 Add a quirk to reset replicas before vbucket move. | |
| Replace unnecessarily complicated misc:letrec-s. | |
| MB-29436 Work around project white-rabbit bugs. | |
| Make sure ns_doctor:wait_statuses/2 returns only fresh statuses. | |
| Return missing nodes from ns_doctor:wait_statuses/2 on timeout. | |
| [backport] Get rid of ns_config:get_global_timeout. | |
| Rework ns_doctor:wait_statuses to return up to date statuses. | |
| Allow starting dcp_replicator on any node. | |
| Bump mb_master advertised version. | |
| Add ns_config:search_node_with_default/4. | |
| Don't require the dcp_consumer_conn to be registered. | |
| Return only live(-ish) children from dcp_sup:get_children. | |
| Conflicts: | |
| include/ns_common.hrl | |
| src/dcp_consumer_conn.erl | |
| src/dcp_replicator.erl | |
| src/dcp_sup.erl | |
| src/ns_cluster.erl | |
| src/ns_config_default.erl | |
| Change-Id: Ia400ec56237851dfc168894c127fb22e838e99b6 | |
| * Commit: a5db776e26b8cd083d2c2d19be395c5a3bc971d2 | |
| MB-29436 Config upgrade to 5.1.1 | |
| Even though we have the upgrade path to 4.6.5 that marks node as | |
| patched against project intact bugs, the customer might already be | |
| running 5.*. Since ns_config upgrades assume linear version history, | |
| we won't go through the upgrade to 4.6.5. Hence, we need to have the | |
| same upgrade path for upgrades from 5.* to 5.1.1. | |
| Change-Id: I2a274c359e11d58efb207e22d9b430b9666218f6 | |
| Reviewed-on: http://review.couchbase.org/93713 | |
| Tested-by: Aliaksey Artamonau <[email protected]> | |
| Reviewed-by: Artem Stemkovski <[email protected]> | |
| * Commit: 3c68f086f88903274471c66db9bf465ae1d2ed71 | |
| MB-29436 Bump mb_master advertised version. | |
| Change-Id: Ica76238a21afa49a66eb43c9359f0b195d68bafc | |
| Reviewed-on: http://review.couchbase.org/93712 | |
| Tested-by: Aliaksey Artamonau <[email protected]> | |
| Reviewed-by: Artem Stemkovski <[email protected]> | |
| * Commit: 064976f3819b37f8660a97df4ac158f1387cec93 | |
| Merge remote-tracking branch 'couchbase/watson' into spock | |
| * couchbase/watson: | |
| MB-29436 Add a quirk to do trivial moves during rebalance. | |
| MB-29436 Add a quirk to reset replicas before vbucket move. | |
| Replace unnecessarily complicated misc:letrec-s. | |
| MB-29436 Work around project white-rabbit bugs. | |
| Make sure ns_doctor:wait_statuses/2 returns only fresh statuses. | |
| Return missing nodes from ns_doctor:wait_statuses/2 on timeout. | |
| [backport] Get rid of ns_config:get_global_timeout. | |
| Rework ns_doctor:wait_statuses to return up to date statuses. | |
| Allow starting dcp_replicator on any node. | |
| Bump mb_master advertised version. | |
| Add ns_config:search_node_with_default/4. | |
| Don't require the dcp_consumer_conn to be registered. | |
| Return only live(-ish) children from dcp_sup:get_children. | |
| Conflicts: | |
| src/cluster_compat_mode.erl | |
| src/dcp_consumer_conn.erl | |
| src/dcp_replicator.erl | |
| src/dcp_sup.erl | |
| src/dcp_upgrade.erl | |
| src/ns_cluster.erl | |
| src/ns_config_default.erl | |
| src/ns_orchestrator.erl | |
| src/ns_single_vbucket_mover.erl | |
| src/ns_vbucket_mover.erl | |
| Change-Id: I04facb70611efc39e777eb8f11c4e1e043a03113 | |
| * Commit: 72444b84db351e19cfa30df2f3dd466662843997 | |
| MB-29436 Add a quirk to do trivial moves during rebalance. | |
| Which in combination with reset_replicas quirk allows rebuilding all | |
| replicas in all buckets. | |
| Change-Id: I66aa1048df7d66e8118116d0ba321ea1a468edd1 | |
| Reviewed-on: http://review.couchbase.org/93660 | |
| Reviewed-by: Artem Stemkovski <[email protected]> | |
| Tested-by: Artem Stemkovski <[email protected]> | |
| * Commit: 151435321b7f72f272cd571da191111230699c6c | |
| MB-29436 Add a quirk to reset replicas before vbucket move. | |
| It's not enabled by default. But it can be enabled manually in case | |
| replicas a compromised due to cursor dropping bug (or any other | |
| reason). | |
| Change-Id: I7da57e928a5443b3961d3f34206e61181c888717 | |
| Reviewed-on: http://review.couchbase.org/93659 | |
| Reviewed-by: Artem Stemkovski <[email protected]> | |
| Tested-by: Artem Stemkovski <[email protected]> | |
| * Commit: cd58333f64f319e051c5ed186f2fc231123ae3d0 | |
| Replace unnecessarily complicated misc:letrec-s. | |
| Change-Id: I8ea1fab3e3abc8d89a4b98d37852c86f7c5da522 | |
| Reviewed-on: http://review.couchbase.org/93658 | |
| Reviewed-by: Artem Stemkovski <[email protected]> | |
| Tested-by: Artem Stemkovski <[email protected]> | |
| * Commit: 524894eb020c27310817860d74779fc0a5211757 | |
| MB-29436 Work around project white-rabbit bugs. | |
| Workarounds are: | |
| - takeovers for the affected nodes is performed on a separate | |
| connection via the orhcestrator node | |
| - vbucket on the old master is explicitly set to state replica prior | |
| to the takeover | |
| The workarounds are applied based on the node version or self-declared | |
| vulnerability (or rather, lack of thereof) to the bugs. | |
| Change-Id: Ied7052d446bf6183d0503583c8f828e424e15803 | |
| Reviewed-on: http://review.couchbase.org/93657 | |
| Reviewed-by: Artem Stemkovski <[email protected]> | |
| Tested-by: Artem Stemkovski <[email protected]> | |
| * Commit: b0895aaeb0a849184dbcd1dbe8cac0c8c7990763 | |
| Make sure ns_doctor:wait_statuses/2 returns only fresh statuses. | |
| To achieve that we need to exclude "down" statuses from the initial | |
| statuses that we get from ns_doctor. | |
| The only current user of this function is the code for memory quotas | |
| validation. Even though technically this change prevents it from | |
| working in some situations where it previously worked, I believe the | |
| new semantic is clearer. It doesn't pretend that we can deal with down | |
| nodes there. And even previously it only worked only in a limit set of | |
| cases where nodes were down. | |
| Change-Id: I38a5563bf03b6ab2403157ee2dfe1c1c8bcde626 | |
| Reviewed-on: http://review.couchbase.org/93656 | |
| Tested-by: Aliaksey Artamonau <[email protected]> | |
| Reviewed-by: Artem Stemkovski <[email protected]> | |
| * Commit: e4e8864a5de64e42f88d1fc5f65db12b8795972b | |
| Return missing nodes from ns_doctor:wait_statuses/2 on timeout. | |
| Change-Id: I7617cf705843fe2a48c135ee37465899c301bd52 | |
| Reviewed-on: http://review.couchbase.org/93655 | |
| Tested-by: Aliaksey Artamonau <[email protected]> | |
| Reviewed-by: Artem Stemkovski <[email protected]> | |
| * Commit: 25804ff19ceee68bd06bd209d7aa2eac639672f8 | |
| [backport] Get rid of ns_config:get_global_timeout. | |
| Instead have get_timeout honor both per node and global values. | |
| Change-Id: Ibe3d4223f4167bc0cff11b0c9105a831b0580b76 | |
| Original-Change-Id: Idd7387645df375207f88865aaa528a69fbd068a8 | |
| Original-Reviewed-on: http://review.couchbase.org/70197 | |
| Original-Tested-by: Aliaksey Artamonau <[email protected]> | |
| Original-Reviewed-by: Artem Stemkovski <[email protected]> | |
| Reviewed-on: http://review.couchbase.org/93654 | |
| Tested-by: Aliaksey Artamonau <[email protected]> | |
| Reviewed-by: Artem Stemkovski <[email protected]> | |
| * Commit: f58b5401f5c8d54ee60420e7383b2923d54637f7 | |
| Rework ns_doctor:wait_statuses to return up to date statuses. | |
| Previously, it would grab statuses once and then wait for the ones | |
| that are missing. But it would not update the statuses it already | |
| got. Which is not a big deal in most cases, but it's also nice to have | |
| the most recent statuses possible in certain situations. | |
| Change-Id: I1f9b6860616ad73f6bf3656c7d515b6b065c0ddc | |
| Reviewed-on: http://review.couchbase.org/93653 | |
| Tested-by: Aliaksey Artamonau <[email protected]> | |
| Reviewed-by: Artem Stemkovski <[email protected]> | |
| * Commit: 1aaf8aef22641def23db5399d26cdaf0abb357a5 | |
| Allow starting dcp_replicator on any node. | |
| Not just the consumer. | |
| Change-Id: I952189a4d644b6eae3477d3c979e398b6f4d0a44 | |
| Reviewed-on: http://review.couchbase.org/93652 | |
| Tested-by: Aliaksey Artamonau <[email protected]> | |
| Reviewed-by: Artem Stemkovski <[email protected]> | |
| * Commit: 9c5df019c60caf10bc2c59c59df47afef4b39974 | |
| Bump mb_master advertised version. | |
| Change-Id: Ifdd3790d8cf8e11af5e1563216064a31264819f4 | |
| Reviewed-on: http://review.couchbase.org/93651 | |
| Tested-by: Aliaksey Artamonau <[email protected]> | |
| Reviewed-by: Artem Stemkovski <[email protected]> | |
| * Commit: 3ecccb44c938e1fa49285453898159d29fcfe889 | |
| Add ns_config:search_node_with_default/4. | |
| Change-Id: I599136f8954abe2eb43184b1ebad22fc06466c5d | |
| Reviewed-on: http://review.couchbase.org/93650 | |
| Tested-by: Aliaksey Artamonau <[email protected]> | |
| Reviewed-by: Artem Stemkovski <[email protected]> | |
| * Commit: 615f22986af6881a7db239072a8d06461df6748e | |
| Don't require the dcp_consumer_conn to be registered. | |
| To get the list of partitions replicated. | |
| Change-Id: I0efb648cf49625c6c9b6b93a942c95ef485ad473 | |
| Reviewed-on: http://review.couchbase.org/93649 | |
| Tested-by: Aliaksey Artamonau <[email protected]> | |
| Reviewed-by: Artem Stemkovski <[email protected]> | |
| * Commit: 66ff0b223e5a86f6dac1ef613921717670d9ff21 | |
| Return only live(-ish) children from dcp_sup:get_children. | |
| Because that's what the callers in fact expect | |
| (e.g. dcp_replication_manager:get_replictor_pid). | |
| Change-Id: Ia56e37ef16c6afc2b833b5ef82cb7d20a71806ab | |
| Reviewed-on: http://review.couchbase.org/93648 | |
| Tested-by: Aliaksey Artamonau <[email protected]> | |
| Reviewed-by: Artem Stemkovski <[email protected]> | |
| * Commit: f9d314961916aaebbf6a0126cd57fec7eb716ede | |
| MB-29350: Disallow cbas and index disk paths... | |
| ...to be sub-directories of the data path. | |
| If allowed then the cbas/index sub-directory will get deleted as | |
| part of the old bucket cleanup process. | |
| Change-Id: If549326f67271825524b19a55eda0d9957dc9420 | |
| Reviewed-on: http://review.couchbase.org/93462 | |
| Tested-by: Ajit Yagaty <[email protected]> | |
| Well-Formed: Build Bot <[email protected]> | |
| Reviewed-by: Artem Stemkovski <[email protected]> | |
| * Commit: 62157a03ea8102d28dd0c426e2ec6d2ad9429c41 | |
| Expand symbolic links while accepting... | |
| ...data, index and cbas paths. | |
| The following are the changes: | |
| [1] If the paths provided have symbolic links in them, we will | |
| resolve them to the corresponding destination path and store | |
| that. | |
| [2] While checking the path, it's possible that a sub-directory | |
| hierarchy is not present. In such cases, we create the path | |
| by appending the remaining path components (non existent | |
| directory hierarchy) to the partially expanded sub-path. | |
| Change-Id: Ia38945949a53bb614677806d8ac81bc2c965b6ed | |
| Reviewed-on: http://review.couchbase.org/93585 | |
| Tested-by: Ajit Yagaty <[email protected]> | |
| Well-Formed: Build Bot <[email protected]> | |
| Reviewed-by: Artem Stemkovski <[email protected]> | |
| * Commit: 7bd2f81ff6c2d4a4c98d06253932b6c64f36ca27 | |
| [misc] Change realpath* APIs to return remaning... | |
| ...path components when an error is encountered. | |
| The realpath* APIs resolve the symlinks in the path. Currently, if an | |
| error is encountered we return back the partially resolved sub-path. | |
| The purpose of this change is to also return back the path components | |
| that were remaning when the error was encountered. | |
| Change-Id: Ibda6486300207f859d491e38e325c9ea25396dde | |
| Reviewed-on: http://review.couchbase.org/93584 | |
| Tested-by: Ajit Yagaty <[email protected]> | |
| Well-Formed: Build Bot <[email protected]> | |
| Reviewed-by: Artem Stemkovski <[email protected]> | |
| * Commit: 4d63fd3b9a0b04baab2cfd886f28f92876a4ed90 | |
| Formatting... | |
| Change-Id: Iaead05c6c2b41f6f39f7b31db0ca5caea16a232d | |
| Reviewed-on: http://review.couchbase.org/93583 | |
| Tested-by: Ajit Yagaty <[email protected]> | |
| Well-Formed: Build Bot <[email protected]> | |
| Reviewed-by: Artem Stemkovski <[email protected]> | |
| * Commit: 1ea2406256327e28fa4d81fb868141699d4baee7 | |
| [Backport] MB-29443: Remove word “Target” from auditing page | |
| - replaced Target with Audit | |
| - updated the small gray text style | |
| Change-Id: I46e3d7a80170259d98681469dfd16ac061263e6c | |
| Reviewed-on: http://review.couchbase.org/93558 | |
| Reviewed-by: Mike Wiederhold <[email protected]> | |
| Tested-by: Mike Wiederhold <[email protected]> | |
| * Commit: a28fea7b1d636fda2b044151fd1751c3b91315a7 | |
| [Backport] Grammar fix for auto-failover tooltip | |
| - was a “the” in the first sentence | |
| Change-Id: I3311d74a5a8100fcc111a0567947a4fa8cc7cb48 | |
| Reviewed-on: http://review.couchbase.org/93557 | |
| Reviewed-by: Mike Wiederhold <[email protected]> | |
| Tested-by: Mike Wiederhold <[email protected]> | |
| * Commit: 81f7760d9183047f50791c6eb3919698af669162 | |
| [Backport] MB-29315: shorten auto-failover tooltip | |
| - edited tooltip text to shorten it | |
| - moved toggle control CSS from query.css to components.css (unrelated) | |
| Change-Id: Id1082379fd5d64f745c12bd26a54212079481a9e | |
| Reviewed-on: http://review.couchbase.org/93556 | |
| Reviewed-by: Mike Wiederhold <[email protected]> | |
| Tested-by: Mike Wiederhold <[email protected]> | |
| * Commit: 91b6eee909de3a27d9bdaa46d5a513db644d999b | |
| [Backport] MB-29358: Hide compaction settings for standard index on EE | |
| Change-Id: I5121ac91222d33a73fae37688cd3a80700f4e06b | |
| Reviewed-on: http://review.couchbase.org/93555 | |
| Reviewed-by: Mike Wiederhold <[email protected]> | |
| Tested-by: Mike Wiederhold <[email protected]> | |
| * Commit: 63467c15ff9b18679a27b5d2266d4724a5614f2d | |
| [Backport] MB-29242: hide max rollback controls for EE | |
| Change-Id: I69c5ba39d698a0b7210b323a0c234e9461b27f96 | |
| Reviewed-on: http://review.couchbase.org/93554 | |
| Reviewed-by: Mike Wiederhold <[email protected]> | |
| Tested-by: Mike Wiederhold <[email protected]> | |
| * Commit: 64acd6ebb7b7c3b9c1965a77a827118417f94ebb | |
| [Backport] MB-29250: clickable server names | |
| - the server name text cancels the click event so as to be copyable | |
| - that is causing some confusion | |
| - I removed stopPropogation from the server name | |
| - I added a new line “name” to server properties in the expanded row and that is copyable | |
| Change-Id: I1dcd9cf90de42b44829be86b6dd54314a5f0058c | |
| Reviewed-on: http://review.couchbase.org/93553 | |
| Reviewed-by: Mike Wiederhold <[email protected]> | |
| Tested-by: Mike Wiederhold <[email protected]> | |
| CHANGELOG for platform | |
| * Commit: e40463a55244f4741e935b2295e2b225191790c9 | |
| Include missing cmake include CheckIncludeFile | |
| The missing include cause failures on some platforms (I spottet | |
| it on my Windows machine) with some versions of CMake. Given | |
| that we use that function we should include the include file | |
| anyway. | |
| Change-Id: I17b2ef90c5ea0de714bce39219e26d903e8913ed | |
| Reviewed-on: http://review.couchbase.org/93680 | |
| Tested-by: Build Bot <[email protected]> | |
| Reviewed-by: Dave Rigby <[email protected]> | |
| Reviewed-by: Tim Bradgate <[email protected]> | |
| CHANGELOG for query | |
| * Commit: a24cc7f2c3c296a0c4a952f4c4c5d838e154e906 | |
| MB-29530. Consider duplicate index keys | |
| Change-Id: Ic4678510ed89cbb55befa09429703310ac2fc158 | |
| Reviewed-on: http://review.couchbase.org/93783 | |
| Reviewed-by: Bingjie Miao <[email protected]> | |
| Tested-by: Sitaram Vemulapalli <[email protected]> | |
| * Commit: f0471f4be6b315f3fcaaca4a3eecdca984b180d4 | |
| MB-29485 set request and run phase correctly | |
| The Run phase time was being set after the results where sent. | |
| Also, the request time was being sent well after the request started being accepted. | |
| Fixed when times are started and set. | |
| Change-Id: I5863b7dcc1cd2914cce2072ad3d060983ee5f6e3 | |
| Reviewed-on: http://review.couchbase.org/93549 | |
| Reviewed-by: Sitaram Vemulapalli <[email protected]> | |
| Tested-by: Marco Greco <[email protected]> | |
| CHANGELOG for testrunner | |
| * Commit: d3b320536f23f502256853e4c883326b9d94eb7a | |
| CBQE-4333: Add test for invalidating timers after docs gets deleted | |
| Change-Id: I09ad23459524488e96c62f4a7656ec78fa594f57 | |
| Reviewed-on: http://review.couchbase.org/93759 | |
| Reviewed-by: Mihir Kamdar <[email protected]> | |
| Tested-by: Balakumaran G <[email protected]> | |
| * Commit: 8bb21710b71031610efc7ff0e44d3485598bc262 | |
| initial attempt at rqg server pool | |
| Change-Id: I8903bb0369ef809a692e3ab11bc2c008b5ef198b | |
| Reviewed-on: http://review.couchbase.org/93789 | |
| Reviewed-by: Arunkumar Senthilnathan <[email protected]> | |
| Tested-by: Korrigan Clark <[email protected]> | |
| * Commit: b07a2f5bb229945d8c926de6f57ebffcfa0dd956 | |
| fixing some rqg jobs | |
| Change-Id: Ic3537547192fe11fd3b831cf50cbf809cd4ef43b | |
| Reviewed-on: http://review.couchbase.org/93788 | |
| Reviewed-by: Arunkumar Senthilnathan <[email protected]> | |
| Tested-by: Korrigan Clark <[email protected]> | |
| * Commit: 6af5a15ea53c7f7430bf4bc9f05ee692066553e0 | |
| CBQE-3069: 1. Do not loop endlessly if mutations_to_index is not zero\n 2. Fix consistency vector test | |
| Change-Id: I01d41316ecad5580947a7b6665dcca958a4eb227 | |
| Reviewed-on: http://review.couchbase.org/93714 | |
| Tested-by: Aruna Piravi <[email protected]> | |
| Reviewed-by: Thuan Nguyen <[email protected]> | |
| * Commit: 95083c87dc5547993a8baa9fa89a653a7a41df25 | |
| CBQE-4614 add test for issue MB-29119 | |
| Change-Id: Ic3bb7013de19222a7946bca8905a5f886606a5b5 | |
| Reviewed-on: http://review.couchbase.org/93642 | |
| Tested-by: Thuan Nguyen <[email protected]> | |
| Reviewed-by: Arunkumar Senthilnathan <[email protected]> | |
| * Commit: 93c1a3cabe66d907a822d30d7784139b1fff7b74 | |
| Don't sleep for 2 secs before rechecking when we have a timeout | |
| We can burn the extra CPU cycles to recheck more often to speed | |
| up the test instead of waiting 2 whole seconds before checking | |
| again when we have a timeout value some time in the future | |
| (and not a number of retries before giving up) | |
| Change-Id: Ib06d485ae2ef6c12f4811edb88e66779e9b988bb | |
| Reviewed-on: http://review.couchbase.org/92256 | |
| Reviewed-by: Dave Rigby <[email protected]> | |
| Tested-by: Trond Norbye <[email protected]> | |
| * Commit: 27fbfcf6d502e8c3b7a42f839f21ae2bbd6b722d | |
| Don't wait for 2 secs for all buckets to delete unless we deleted any bucket | |
| Change-Id: Ifcf83abdde36d4966a489c97441ea68e198c0225 | |
| Reviewed-on: http://review.couchbase.org/92255 | |
| Reviewed-by: Dave Rigby <[email protected]> | |
| Tested-by: Trond Norbye <[email protected]> | |
| * Commit: 8cf270939e39b994c2cc02e236b47b4eb38dadfc | |
| CBQE-4614 add cmd_set test to check cmd_set number | |
| should increase if there is a set operation. | |
| Change-Id: I82031bbb3e5699f1b3007808e7b2be75b1bff2b4 | |
| Reviewed-on: http://review.couchbase.org/93630 | |
| Tested-by: Thuan Nguyen <[email protected]> | |
| Reviewed-by: Korrigan Clark <[email protected]> | |
| * Commit: b349e7591affbcf15ce1cc5fceb1ab81f0e3639b | |
| CBQE-4614 add test role data_monitoring should be allowed to run mctimings | |
| Change-Id: I75eef1d702cf61ef606d9c5f41b6a5361eb1eae7 | |
| Reviewed-on: http://review.couchbase.org/93621 | |
| Tested-by: Thuan Nguyen <[email protected]> | |
| Reviewed-by: Arunkumar Senthilnathan <[email protected]> | |
| * Commit: 6cdf08d7cf1217398d52fd615cc61b8bc98009ff | |
| CBQE-4614 bypass check build number format | |
| Change-Id: I3c24ed7313245fb5191bb36c0925ee3c9a7b3038 | |
| Reviewed-on: http://review.couchbase.org/93579 | |
| Tested-by: Thuan Nguyen <[email protected]> | |
| Reviewed-by: Aruna Piravi <[email protected]> | |
| CHANGELOG for tlm | |
| * Commit: 86f2ed0b1fc8b1edd1897be16a95e76eb00d97cf | |
| CBD-2303: Include building and installing m4 for flex | |
| Change-Id: I3d4ed2f9dd2e967bd2b7e907796a784f3a4de701 | |
| Reviewed-on: http://review.couchbase.org/93735 | |
| Reviewed-by: Chris Hillery <[email protected]> | |
| Tested-by: Kenneth Lareau <[email protected]> | |
| Tested-by: Chris Hillery <[email protected]> | |
| Couchbase build team - [email protected] |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment