apple/foundationdb
1
0
Fork 0
mirror of https://github.com/apple/foundationdb.git synced 2025-05-22 23:19:02 +08:00
Code Issues Packages Projects Releases Wiki Activity
14,326 Commits 35 Branches 318 Tags
df90cc89de
Commit Graph

80 Commits

Author SHA1 Message Date
FDB Formatster
df90cc89de apply clang-format to *.c, *.cpp, *.h, *.hpp files 2021-03-10 10:18:07 -08:00
Andrew Noyes
cec050a1f8 Apply clang-tidy's modernize-use-override fix 2021-01-26 02:04:12 +00:00
sfc-gh-tclinkenbeard
4669f837fa Add uses of makeReference 2020-11-07 22:10:18 -08:00
sfc-gh-tclinkenbeard
035708f43a Post-merge cleanup 2020-10-29 22:35:27 -07:00
sfc-gh-tclinkenbeard
cf4c8e375f Merge remote-tracking branch 'origin/release-6.3' into merge 2020-10-29 22:15:41 -07:00
Jingyu Zhou
e183a10c8c Add "waitForDestUID" flag for abortBackup 2020-10-20 22:55:31 -07:00
Jingyu Zhou
47a9bc9d3d Fix commit_unknown_result issue 2020-10-18 23:11:19 -07:00
Jingyu Zhou
bfd3328448 Fix a race between submit and abort backup 
After submit a backup, immediately abort the backup may cause a rare race
condition, which results in BackupCorrectnessLeftoverVersionKey error.

Specifically, in the StartFullBackupTaskFunc:
1st Txn sets the destUid at the source database and the 2nd Txn writes the dest
DB.

An abort can come after the 1st Txn succeeds, and clears the config rage so
that the 2nd Txn above would fail. Because 2nd Txn didn't write destUid, the
3rd Txn of abort can't read the correct source DB for latestVersionKey, which
contains the destUid value.

The fix is to let the 1st Txn of abort to wait until destUid becomes valid.
 2020-10-18 23:11:15 -07:00
sfc-gh-tclinkenbeard
8ae0679112 Remove unnecessary separation of declaration and definition of version field 2020-10-04 23:27:05 -07:00
sfc-gh-tclinkenbeard
d13a9d3c91 Make EBackupState an enum class 2020-09-23 15:14:09 -07:00
Jingyu Zhou
90654628fb
Merge pull request from apple/release-6.3 
Merge Release 6.3
 2020-07-10 12:54:37 -07:00
Oleg Samarin
4c9df78076 dstonly 2020-07-10 15:13:42 +03:00
Oleg Samarin
6d70887d68
Update fdbclient/DatabaseBackupAgent.actor.cpp 
Speed up sync between Primary and DR: Event Name Correction

Co-authored-by: Jingyu Zhou <jingyuzhou@gmail.com>
 2020-07-04 21:37:28 +03:00
Oleg Samarin
6c23901e61 Speed up sync between Primary and DR 2020-07-03 11:38:54 +03:00
Oleg Samarin
1832b88e3f Removed debug messages 2020-06-26 13:34:52 +03:00
Oleg Samarin
d887d5dd70 Increased CopyLogRange performance 2020-06-26 13:34:35 +03:00
Evan Tschannen
ced65cd30b finished explicitly versioning everything stored in the database 2020-05-22 17:14:21 -07:00
Evan Tschannen
0420b3e786 fix compile error 2020-04-29 14:05:53 -07:00
Evan Tschannen
7cebe743f9 A number of bug fixes of rare correctness errors 2020-04-29 13:50:13 -07:00
Evan Tschannen
a486ec2de0 pipelined fdbdr status 2020-02-25 15:48:00 -08:00
Evan Tschannen
daee15cbb5 fix: starting a DR should do the commit on the first proxy to ensure all mutations from previous backups have been flushed 2020-02-25 12:35:24 -08:00
Evan Tschannen
73ad702d14 Clients which fetch status should not disconnect from the coordinators and cluster controller between each retrieval 2020-01-22 15:41:22 -08:00
Steve Atherton
4ff058e86b Backup and DR layer status document generation now uses snapshot reads for all keys read to avoid unnecessary conflicts when read during a status update or cleanup transaction. Since many of the keys read use wrapper functions, all of the underlying functions in BackupAgentBase and its two implementations also required a snapshot mode argument. All snapshot arguments default to false to match the underlying FDB API get/getrange methods. 2019-12-19 00:29:35 -08:00
Evan Tschannen
628b4e0220 added a warning if multiple log ranges exist for the same range 2019-10-02 17:06:19 -07:00
Evan Tschannen
9ec9f41d34 backup and DR would not share mutations if they were started on different versions of FDB 2019-10-01 18:52:07 -07:00
Evan Tschannen
ef01ad2ed8 optimized log range clearing to clear everything for each possible hash (256 clears) if that would be more efficient than one clear per second that has elapsed 
aborting a DR without the —cleanup flag will still attempt to cleanup for 30 seconds before giving up
added a cleanup command to fdbbackup which can remove mutations from orphaned DRs which were stopped without the —cleanup flag
 2019-09-27 18:32:27 -07:00
Jingyu Zhou
3a63d053e9 Address review comments for PR#1725 2019-06-20 14:06:32 -07:00
A.J. Beamon
5f55f3f613 Replace g_random and g_nondeterministic_random with functions deterministicRandom() and nondeterministicRandom() that return thread_local random number generators. Delete g_debug_random and trace_random. Allow only deterministicRandom() to be seeded, and require it to be seeded from each thread on which it is used. 2019-05-10 14:01:52 -07:00
Andrew Noyes
781b6ece77 Fix OPEN_FOR_IDE -Wunused-variable warnings 
CC ,
2019-04-16 15:28:01 -07:00
mpilman
1c16f87a4e Remove trace-calls to printable (in non-workloads) 2019-04-05 13:12:19 -07:00
Steve Atherton
8aab719c22
Merge branch 'master' into feature-backup-json 2019-03-12 18:23:16 -07:00
Stephen Atherton
023bbb566f Renamed backup state enums for clarity, added backup state names. Changed Epochs to EpochSeconds in backup JSON along with some other renaming/moving of fields, and added information about snapshot dispatch. Changed timestamp format for input/output in all backup/restore contexts to be a fully qualified time with timezone offset. Added information about the last snapshot dispatch to backup config and status (not yet populated). 2019-03-10 16:00:01 -07:00
Evan Tschannen
f1897f3eb6 Merge branch 'master' into feature-metadata-version 
# Conflicts:
#	fdbclient/NativeAPI.actor.cpp
 2019-03-04 21:06:16 -08:00
Balachandar Namasivayam
a258df32f6 Skip switchover checks for force option. 2019-03-04 15:58:36 -08:00
Balachandar Namasivayam
7cf71b0931 Add some basic checks before doing an atomic switchover. 2019-03-01 14:49:04 -08:00
Evan Tschannen
3da85f3acd implemented the \xff/metadataVersion key, which can be used by layers to help them cheaply cache metadata and know when their cache is invalid 2019-02-28 17:45:00 -08:00
mpilman
c1592b4c3a Several minor fixes within fdbclient 2019-02-19 15:16:59 -08:00
mpilman
78dd80ea8a Proper fwd decl in BackupAgent 
Also BackupAgent.h -> BackupAgent.actor.h
 2019-02-19 15:16:59 -08:00
mpilman
3cb2391b58 use proper fwd declarations in ManagementAPI 
Also ManagementAPI.h -> ManagementAPI.actor.h
 2019-02-19 15:16:59 -08:00
mpilman
479a4d7c22 Minor fixes in fdbclient for intellisense 2019-02-19 15:16:59 -08:00
Andrew Noyes
067a445e06 Replace unused _ variables with wait(success(...)) 2019-02-12 17:30:30 -08:00
A.J. Beamon
d4d5740282 * Add Optional.map and ErrorOr.map. 
* Rename Optional/ErrorOr cast_to to castTo.
* Make printable(Optional<T>) templated rather than restricted to StringRef types.
* Fixes bug in (unused) ErrorOr.castTo where an ErrorOr that was not set would lose its error.
 2019-01-11 09:03:38 -08:00
Robert Escriva
268093a96d Adjust all includes to be relative to the root. 
Remove the use of relative paths.  A header at foo/bar.h could be included by
files under foo/ with "bar.h", but would be included everywhere else as
"foo/bar.h".  Adjust so that every include references such a header with the
latter form.

Signed-off-by: Robert Escriva <rescriva@dropbox.com>
 2018-10-19 17:35:33 +00:00
Alex Miller
535b5701e5 Rewrite all Void _ = wait(...) -> wait(...). 
This takes advantage of the new actorcompiler functionality to avoid
having duplicate definitions of `Void _` when trying to feed the
un-actorompiled source through clang.
 2018-08-14 15:50:26 -07:00
A.J. Beamon
99c9958db7 Some more trace event normalization 2018-06-08 13:57:00 -07:00
A.J. Beamon
e5488419cc Attempt to normalize trace events: 
* Detail names now all start with an uppercase character and contain no underscores. Ideally these should be head-first camel case, though that was harder to check.
* Type names have the same rules, except they allow one underscore (to support a usage pattern Context_Type). The first character after the underscore is also uppercase.
* Use seconds instead of milliseconds in details.

Added a check when events are logged in simulation that logs a message to stderr if the first two rules above aren't followed.

This probably doesn't address every instance of the above problems, but all of the events I was able to hit in simulation pass the check.
 2018-06-08 11:11:08 -07:00
Evan Tschannen
e82985aea2 fix: continue setting beginVersion so that versions between 5.2.0 and 5.2.2 do not crash when decoding tasks created by 5.2.3 2018-06-06 13:34:22 -07:00
Evan Tschannen
4120062bb9 fix: backup initialized its begin version at 1 instead of the read version of the starting transaction 
fix: erasing log ranges did not properly divide up work between transactions to prevent making transactions which were too large
 2018-06-06 13:05:53 -07:00
Evan Tschannen
8930c2e3db DR upgrade tests now test the durability of the data. 2018-05-09 15:11:05 -07:00
Yichi Chiang
c721ab6854 Fix review comments 2018-04-27 13:54:34 -07:00