Vladislav
d9f8f2553b
chore: fix return on bad status ( #3758 )
2024-09-22 01:36:39 +03:00
Roman Gershman
cce2eb35ed
chore: refactor a lambda function into a named one ( #3753 )
...
Also did some cosmetic improvements. No functionality should be changed.
Signed-off-by: Roman Gershman <roman@dragonflydb.io>
2024-09-22 01:35:56 +03:00
Andy Dunstall
9dd79657ce
fix: zset store conclude transaction on error ( #3755 )
2024-09-21 19:08:53 +03:00
Borys
ce79da0f7a
fix: add value range check for SETBIT command ( #3750 )
2024-09-20 18:20:35 +03:00
Roman Gershman
c9a2334f6d
fix: allow the healthcheck run in non-privileged containers as well ( #3731 )
...
fix: allow the healthcheck running in non-privileged containers as well
Fixes #3644 (again).
Signed-off-by: Roman Gershman <roman@dragonflydb.io>
2024-09-20 05:41:06 +00:00
Kostas Kyrimis
ed21867fe9
chore: add missing await in test_take_over_seeder ( #3744 )
...
* add missing await
Signed-off-by: kostas <kostas@dragonflydb.io>
2024-09-19 17:03:11 +00:00
Roman Gershman
abf3acec4a
chore: introduce a Clone function for the dense set ( #3740 )
...
* chore: introduce a Clone function for the dense set
We use a state machine to prefetch data in batches.
After this change, the hot spots are predominantly inside ObjectClone and
Hash methods.
All in all benchmarks show ~45% CPU reduction:
```
BM_Clone/elements:32000 1517322 ns 1517338 ns 2772
BM_Fill/elements:32000 841087 ns 841097 ns 4900
```
---------
Signed-off-by: Roman Gershman <roman@dragonflydb.io>
2024-09-19 16:14:33 +03:00
Vladislav
3af2dfc4e7
chore: add SetReplies ( #3727 )
2024-09-19 12:54:25 +03:00
Kostas Kyrimis
0e0b2e78a4
chore: change log level to warning for empty keys ( #3722 )
...
* adjust log level to warning for allowed empty keys in rdb_load and rdb_save
Signed-off-by: kostas <kostas@dragonflydb.io>
2024-09-19 09:45:09 +00:00
Shahar Mike
55e3647248
chore: Switch ports for cluster_mgr_test.py ( #3741 )
...
We saw failures due to port already in use
2024-09-19 12:32:31 +03:00
adiholden
409c2a3beb
test: add test for replication deadlock on replication timeout ( #3691 )
...
* test: add test for replication deadlock on replication timeout
Signed-off-by: adi_holden <adi@dragonflydb.io>
2024-09-19 12:11:28 +03:00
Borys
efa4efd2bf
refactor: use CmdArgParser for XGROUP command ( #3739 )
2024-09-18 22:30:37 +03:00
Borys
bbaa2669f9
test: unskip test for debugging purpose ( #3738 )
2024-09-18 14:13:07 +00:00
Borys
f122a19a02
test: add tests for replication ( #3734 )
...
* test: add tests for replication
2024-09-18 16:32:21 +03:00
Kostas Kyrimis
6e45c9c3e2
fix: properly track json memory usage ( #3641 )
...
* add JsonMemTracker
* add logic based on MiMallocResource deltas that calculates json object usage
* add test
Signed-off-by: kostas <kostas@dragonflydb.io>
2024-09-18 13:08:43 +00:00
Stepan Bagritsevich
b235617a0d
fix(json_family): Fix out of bound ranges for the JSON.ARR* commands ( #3712 )
...
* fix(json_family): Fix out of bound ranges for theJSON.ARR* commands
Signed-off-by: Stepan Bagritsevich <stefan@dragonflydb.io>
* refactor(json_family): address comments
Signed-off-by: Stepan Bagritsevich <stefan@dragonflydb.io>
* refactor(json_family): address comments 2
Signed-off-by: Stepan Bagritsevich <stefan@dragonflydb.io>
---------
Signed-off-by: Stepan Bagritsevich <stefan@dragonflydb.io>
2024-09-18 14:31:17 +02:00
Vladislav
41ba864924
chore: Remove ReqSerializer ( #3721 )
...
Signed-off-by: Vladislav <vladislav.oleshko@gmail.com>
2024-09-18 14:31:47 +03:00
Shahar Mike
ffb4c2b601
fix: Fix test_take_over_seeder ( #3733 )
...
The test assumed any shutdown will take not more than 1s. This doesn't
always hold, and also waiting for 1s isn't ideal because usually it
takes less than that.
Changed to use `assert_eventually` instead.
Fixes #3684
2024-09-18 14:17:09 +03:00
Shahar Mike
1c6be62a0b
fix: Fix cluster_mgr.py ( #3730 )
...
We updated the reply of `SLOT-MIGRATION-STATUS`, so `cluster_mgr.py`
needs to be adjusted as well.
2024-09-18 11:44:15 +03:00
Shahar Mike
a115bc2b9f
fix: Fix test test_client_pause_with_replica ( #3729 )
...
There are 2 minor issues with this test:
1. It specified `cmdstat_replconf` as `cmd_stats` instead of `cmd`,
that's clearly a typo as `cmd_stats` is a map with stats, while
`replconf` is a Dragonfly command
2. Command `MULTI` is allowed to run even when the server is in paused
state, see
[here](https://github.com/dragonflydb/dragonfly/blob/main/src/server/main_service.cc#L1197 ):
```
// Don't interrupt running multi commands or admin connections.
```
Fixes #3675
2024-09-18 09:40:26 +03:00
Stepan Bagritsevich
ae5ce9b497
fix(json_family): Separate double and int values during the comparison of the JSON objects ( #3711 )
...
* fix(json_family): Separate the double and int values in JSON commands
Signed-off-by: Stepan Bagritsevich <stefan@dragonflydb.io>
* refactor(json_family): Address comments
Signed-off-by: Stepan Bagritsevich <stefan@dragonflydb.io>
---------
Signed-off-by: Stepan Bagritsevich <stefan@dragonflydb.io>
2024-09-18 07:24:48 +02:00
Stepan Bagritsevich
824af02f6f
fix(json_family): Fix JSON.ARRPOP command in legacy mode should not return WRONGTYPE error ( #3683 )
...
* fix(json_family): Fix WRONGTYPE error for the JSON legacy mode in the JSON.ARRPOP command
Signed-off-by: Stepan Bagritsevich <stefan@dragonflydb.io>
* refactor(json_family): address comments
Signed-off-by: Stepan Bagritsevich <stefan@dragonflydb.io>
* refactor(json_family): address comments 2
Signed-off-by: Stepan Bagritsevich <stefan@dragonflydb.io>
---------
Signed-off-by: Stepan Bagritsevich <stefan@dragonflydb.io>
2024-09-18 07:24:18 +02:00
Andy Dunstall
a64fc74ce1
tests: fix and enable s3 snapshot test ( #3720 )
...
* test: fix s3 snapshot test
* ci: configure s3 regression test
* tests: only run s3 snapshot test if bucket not empty
2024-09-17 17:35:53 +03:00
Kostas Kyrimis
8a34b3e730
chore: enable ReplyGuard in ReplyBuilder2 ( #3705 )
...
* add ReplyGuard in ReplyBuilder2
Signed-off-by: kostas <kostas@dragonflydb.io>
2024-09-17 13:37:23 +03:00
Kostas Kyrimis
6f84115152
chore: add log info on failed commands ( #3694 )
...
* log errors on failed commands
Signed-off-by: kostas <kostas@dragonflydb.io>
2024-09-17 13:07:46 +03:00
Shahar Mike
51746d99c7
fix(cluster): Do not Pause() replication / migrations ( #3716 )
...
Pre-this change, whenever Dragonfly was paused (either by a user or by
internal processes like takeover or slot migration finalization),
migrations and replications were also paused.
This could cause timing issues, which sometime result in migration
failures. Specifically, when 2 nodes have migrations from one to the
other **in parallel** (A->B and B->A), the `Pause()` that happens on A
(which happens because it's a source node) will stop it from processing
incoming traffic from B (incoming because it is also a target node).
If timed correctly, it will be locked until it times out, and so the
migration will fail.
The fix is to prevent replications and migrations from adhering to
`Pause()`s, which I think should not have happened in the first place
because they should use the admin port anyway.
Fixes #3319
2024-09-17 10:47:55 +03:00
Andy Dunstall
b9ff6934e8
fix: fix s3 load snapshot ( #3717 )
2024-09-17 07:17:24 +01:00
romange
6f3da56e75
chore(helm-chart): update to v1.22.2
2024-09-16 20:06:02 +00:00
Roman Gershman
e21ba0b3d9
chore: symbolize stack traces in tests upon crash ( #3714 )
...
We disable address space randomization when building the binary
and use addr2line to symbolize the stacktrace if it exists.
Signed-off-by: Roman Gershman <roman@dragonflydb.io>
2024-09-16 13:43:16 +03:00
Borys
e2a852b7e4
fix: add default value has_mc_flag field ( #3710 )
2024-09-16 10:03:08 +03:00
Roman Gershman
267bd431e4
chore: add clone benchmark ( #3709 )
2024-09-15 13:10:43 +03:00
Borys
93de559977
Update dflycluster slot-migration-status reply ( #3707 )
...
* feat: update DFLYCLUSTER SLOT-MIGRATION-STATUS reply
2024-09-15 09:44:40 +03:00
Kostas Kyrimis
b5929f0162
fix: allow parsing extra spaces on acl files ( #3703 )
...
* allow parsing extra whitespace characters in acl files
Signed-off-by: kostas <kostas@dragonflydb.io>
2024-09-13 10:17:20 +03:00
Yang Hau
35c70dbfe1
feat(core): Support RISCV RVV ( #3655 )
2024-09-12 18:40:46 +03:00
Kostas Kyrimis
4988831b6f
chore: pull latest helio ( #3700 )
...
Signed-off-by: kostas <kostas@dragonflydb.io>
2024-09-12 15:51:23 +03:00
Stepan Bagritsevich
3815cda26d
fix(json_family) Add NOESCAPE option to the JSON.GET command ( #3685 )
...
* fix(json_family) Add NOESCAPE option to the JSON.GET command
Signed-off-by: Stepan Bagritsevich <stefan@dragonflydb.io>
* refactor(json_family): address comments
Signed-off-by: Stepan Bagritsevich <stefan@dragonflydb.io>
* refactor(json_family): address comments 2
Signed-off-by: Stepan Bagritsevich <stefan@dragonflydb.io>
---------
Signed-off-by: Stepan Bagritsevich <stefan@dragonflydb.io>
2024-09-12 13:04:46 +02:00
Kostas Kyrimis
5819755af1
fix: test_replicaof_reject_on_load assert failure ( #3697 )
...
* increase snapshot size for the test
Signed-off-by: kostas <kostas@dragonflydb.io>
2024-09-12 09:57:48 +03:00
Borys
bae2767707
test: fix test_cluster_replication_migration ( #3699 )
2024-09-11 23:00:53 +03:00
Kostas Kyrimis
d041386184
fix: test_acl_revoke_pub_sub_while_subscribed ( #3680 )
...
* add logs
* add asyncio sleep to avoid producer stalls
Signed-off-by: kostas <kostas@dragonflydb.io>
2024-09-11 12:00:49 +03:00
Stepan Bagritsevich
3ece1725a1
fix(json_family): Fix the JSON.SET bug if the path is in legacy mode and is not the root ( #3693 )
...
Signed-off-by: Stepan Bagritsevich <stefan@dragonflydb.io>
2024-09-10 15:50:32 +02:00
Borys
9b8aa8eab4
fix: join for cancel incoming migration ( #3692 )
2024-09-10 14:42:37 +03:00
adiholden
e71f679386
fix(server): fix replication master deadlock on cancelation flow ( #3686 )
...
* fix server: fix replication deadlock on cancelation
Signed-off-by: adi_holden <adi@dragonflydb.io>
2024-09-10 14:13:38 +03:00
Roman Gershman
bdc578acef
chore: limit number of descriptors in the exec map ( #3688 )
...
For some cases, this map can grow indefinitely.
This change makes it less detailed by makes sure that number of possible keys is bounded.
Still it can provide a good summary of nature of exec transactions.
Signed-off-by: Roman Gershman <roman@dragonflydb.io>
2024-09-10 07:50:30 +00:00
Roman Gershman
3cdc8fa128
chore: add a script that parses allocator tracking logs ( #3687 )
2024-09-10 07:26:44 +00:00
Roman Gershman
257749263b
chore: adjust RdbChannel sizes ( #3676 )
...
Fixes #3658
Signed-off-by: Roman Gershman <roman@dragonflydb.io>
2024-09-09 22:21:42 +03:00
Borys
35c287b813
test: unskip cluster tests and add debug info ( #3681 )
2024-09-09 22:21:17 +03:00
Stepan Bagritsevich
2fad54e41f
fix(search_family): Fix LOAD option behavior in the FT.AGGREGATE command ( #3660 )
...
* fix(search_family): Fix LOAD option behavior in the FT.AGGREGATE command
fixes dragonflydb#3646
Signed-off-by: Stepan Bagritsevich <stefan@dragonflydb.io>
* Update src/server/search/search_family.cc
Co-authored-by: Shahar Mike <chakaz@users.noreply.github.com>
Signed-off-by: Stepan Bagritsevich <43710058+BagritsevichStepan@users.noreply.github.com>
---------
Signed-off-by: Stepan Bagritsevich <stefan@dragonflydb.io>
Signed-off-by: Stepan Bagritsevich <43710058+BagritsevichStepan@users.noreply.github.com>
Co-authored-by: Shahar Mike <chakaz@users.noreply.github.com>
2024-09-09 09:21:48 +00:00
adiholden
c34f2b7eeb
server logs: change script error to warning ( #3670 )
...
Signed-off-by: adi_holden <adi@dragonflydb.io>
2024-09-09 11:23:37 +03:00
Shahar Mike
b10a4a5348
feat(server): Support CLIENT SETINFO ( #3673 )
...
Add support for `CLIENT SETINFO <LIB-NAME | LIB-VER>` and also return
that as part of `CLIENT LIST`, like Valkey.
Fixes #3137
2024-09-09 11:03:05 +03:00
Roman Gershman
b7b96424e4
deprecate RecordsPopper and serialize channel records during push ( #3667 )
...
chore: deprecate RecordsPopper and serialize channel records during push
Records channel is redundant for DFS/replication because we have single producer/consumer
scenario and both running on the same thread. Unfortunately we need it for RDB snapshotting.
For non-rdb cases we could just pass a io sink to the snapshot producer,
so that it would use it directly instead of StringFile inside FlushChannelRecord.
This would reduce memory usage, eliminate yet another memory copy and generally would make everything simpler.
For that to work, we must serialize the order of FlushChannelRecord, and that's implemented by
this PR. Also fixes #3658 .
Signed-off-by: Roman Gershman <roman@dragonflydb.io>
2024-09-09 06:19:04 +00:00
