1. taskQueueCapacity 256 is too small for production when we want to
re-write the entire collection
2. tasks should be cleaned when unable to recover, or the meta will
remain in etcd forever later.
Signed-off-by: yangxuan <xuan.yang@zilliz.com>
issue: #33550
balance segment and balance segment execute at same time, which will
cause bounch of corner case.
This PR disable simultaneous balance of segments and channels
Signed-off-by: Wei Liu <wei.liu@zilliz.com>
issue: #37830
casue dist handler doesn't set channel's version, so if channel checker
try to dedup channel, it may release the new delegator after balance
finished.
this PR fix the way to set proper version for channel.
Signed-off-by: Wei Liu <wei.liu@zilliz.com>
issue: #37851
- make a global thread pool at tantivy temporally.
- set 1 but not 4 threads for inverted text index.
Signed-off-by: chyezh <chyezh@outlook.com>
issue : https://github.com/milvus-io/milvus/issues/36864
I have a few questions regarding my approach.I will consolidate them
here for feedback and review.Thanks
---------
Signed-off-by: Nischay Yadav <nischay.yadav@ibm.com>
Signed-off-by: Nischay <Nischay.Yadav@ibm.com>
Custom `jsonRender` that encodes JSON data directly into the response
stream, it uses less memory since it does not buffer the entire JSON
structure before sending it, unlike `c.JSON` in `HTTPReturn`, which
serializes the JSON fully in memory before writing it to the response.
Benchmark testing shows that, using the custom render incurs no
performance loss and reduces memory consumption by nearly 50%.
issue: https://github.com/milvus-io/milvus/issues/37671
---------
Signed-off-by: bigsheeper <yihao.dai@zilliz.com>
issue: #37718
This PR refine the shard client ref counter, dec ref counter won't
release client anymore, and only permit shard client manager to remove
client.
---------
Signed-off-by: Wei Liu <wei.liu@zilliz.com>
Limit the maximum concurrency of channel tasks for each DataNode to
prevent excessive subscriptions from causing DataNode OOM.
issue: https://github.com/milvus-io/milvus/issues/37665
Signed-off-by: bigsheeper <yihao.dai@zilliz.com>
When there are a large number of segments, the metrics consume a lot of
memory. This PR Remove segment-level tag from monitoring metrics.
issue: https://github.com/milvus-io/milvus/issues/37636
Signed-off-by: bigsheeper <yihao.dai@zilliz.com>
issue: #36621
* The front-end code of the management interface has been refactored
using React.
* The interaction and display of the front-end interface have been
optimized.
Signed-off-by: mimo.oyn <ning.ouyang@zilliz.com>
Co-authored-by: zilliz <zilliz@zillizdeMacBook-Pro.local>
issue: #37679
pr #36549 introduce the logic error which update current target when
only parts of channel is ready.
This PR fix the logic error and let dist handler keep pull distribution
on querynode until all delegator becomes serviceable.
Signed-off-by: Wei Liu <wei.liu@zilliz.com>
issue: #37579
If the schema includes large varchar fields, a few thousand rows can
reach hundreds of MB in size. Therefore, if the batch size of the
segment writer is large, it will produce relatively large `binlogs`,
which can cause datanode to run out of memory (OOM) during compaction.
Signed-off-by: Cai Zhang <cai.zhang@zilliz.com>
issue: #37665#37631#37620#37587#36906
knowhere has add default nlist value, so some invalid param test ut with
no nlist param will be valid.
Signed-off-by: xianliang.li <xianliang.li@zilliz.com>
issue: ##36621
- For simple types in a struct, add "string" to the JSON tag for
automatic string conversion during JSON encoding.
- For complex types in a struct, replace "int64" with "string."
Signed-off-by: jaime <yun.zhang@zilliz.com>
issue: #37640
fix the pr #36549
cause balance channel will wait until new delegator becomes serviceable,
but new delegator need to sync target version then becomes serviceable,
and sync target version need to be wait all replica load done. so if
increasing replica number and balance channel happens at same time,
logic dead lock occurs.
Signed-off-by: Wei Liu <wei.liu@zilliz.com>
issue: #36293#36242
after qn recover, delegator may be loaded in new node, after all segment
has been loaded, delegator becomes serviceable. but delegator's target
version hasn't been synced, and if search/query comes, delegator will
use wrong target version to filter out a empty segment list, which
caused empty search result.
This pr will block delegator's serviceable status until target version
is synced
---------
Signed-off-by: Wei Liu <wei.liu@zilliz.com>
issue: #37115
the old implementation update shard cache and shard client manager at
same time, which causes lots of conor case due to concurrent issue
without lock.
This PR decouple shard client manager from shard cache, so only shard
cache will be updated if delegator changes. and make sure shard client
manager will always return the right client, and create a new client if
not exist. in case of client leak, shard client manager will purge
client in async for every 10 minutes.
---------
Signed-off-by: Wei Liu <wei.liu@zilliz.com>
issue: #33285
- Modify the proto of consumer of streaming service.
- Make VChannel as a required option for streaming
---------
Signed-off-by: chyezh <chyezh@outlook.com>
Related to #35415
In rolling upgrade, legacy proxy may dispatch load request wit empty
load field list. The upgraded querycoord may report error by mistake
that load field list is changed.
This PR:
- Auto field empty load field list with all user field ids
- Refine the error messag when load field list updates
- Refine load job unit test with service cases
Signed-off-by: Congqi Xia <congqi.xia@zilliz.com>
When there're a lot of loaded collections, they would occupy the target
observer scheduler’s pool. This prevents loading collections from
updating the current target in time, slowing down the load process.
This PR adds a separate target dispatcher for loading collections.
issue: https://github.com/milvus-io/milvus/issues/37166
---------
Signed-off-by: bigsheeper <yihao.dai@zilliz.com>
Close (unsubscribe) the msg stream after completing the
`PreCreatedTopic` check to prevent backlog issue.
issue: https://github.com/milvus-io/milvus/issues/36021
Signed-off-by: bigsheeper <yihao.dai@zilliz.com>
Since master branch run cpp unittest in jenken, the cpp unit test
building could be skipped since the output is not needed in github
runner.
Signed-off-by: Congqi Xia <congqi.xia@zilliz.com>
When check health logic failed to collection not-queryable, the related
reason is hard to find in log.
This PR add context for log with trace id and print unqueryable
collection info log.
Signed-off-by: Congqi Xia <congqi.xia@zilliz.com>
issue: #36977
with node_label_filter on resource group, user can add label on
querynode with env `MILVUS_COMPONENT_LABEL`, then resource group will
prefer to accept node which match it's node_label_filter.
then querynode's can't be group by labels, and put querynodes with same
label to same resource groups.
Signed-off-by: Wei Liu <wei.liu@zilliz.com>
issue: #37172
- add redo interceptor to implement append context refresh. (make new
timetick)
- add create segment handler for flusher.
- make empty segment flushable and directly change it into dropped.
- add create segment message into wal when creating new growing segment.
- make the insert operation into following seq: createSegment -> insert
-> insert -> flushSegment.
- make manual flush into following seq: flushTs -> flushsegment ->
flushsegment -> manualflush.
---------
Signed-off-by: chyezh <chyezh@outlook.com>
issue: #33744
1. Segments generated from inserts will be loaded as growing until they
are sorted by primary key.
2. This PR may increase memory pressure on the delegator, but we need to
test the performance of stats. In local testing, the speed of stats is
greater than the insert speed.
Signed-off-by: Cai Zhang <cai.zhang@zilliz.com>
issue: #37166
cause the misuse of timer.Reset, which cause dispatcher failed to send
msg to virtual channel buffer, and dispatcher do splitting again and
again, which hold the dispatcher manager's lock, block watching channel
progress.
This PR fix the misuse of timer.Reset
Signed-off-by: Wei Liu <wei.liu@zilliz.com>
Make the parameter input method consistent with miluvs-client.
Signed-off-by: lixinguo <xinguo.li@zilliz.com>
Co-authored-by: lixinguo <xinguo.li@zilliz.com>
issue: #36858
- Start channel manager on datacoord, but with empty assign policy in
streaming service.
- Make collection at dropping state can be recovered by flusher to make
sure that
milvus consume the dropCollection message.
- Add backoff for flusher lifetime.
- remove the proxy watcher from timetick at rootcoord in streaming
service.
Also see the better fixup: #37176
---------
Signed-off-by: chyezh <chyezh@outlook.com>
Cgo API cost is not observerable since not metrics is related to them.
This PR add metrics for some sync cgo call related to load & write
---------
Signed-off-by: Congqi Xia <congqi.xia@zilliz.com>
Previously failed label is used for canceled storage op, which may cause
wrong alarm when user cancel load operation or etc. This PR utilizes
cancel label when such case happens.
Signed-off-by: Congqi Xia <congqi.xia@zilliz.com>
issue: #36970
cause release segment and balance channel may happen at same time, and
before new delegator become serviceable, if release segment exeuctes on
new delegator, and search/query comes on old delegator, then release
segment and query segment happens in parallel, if release segment
execute first in worker, then search/query will got a SegmentNodeLoaded
error.
This PR add serviceable filter on delegator, then all load/release
segment operation will happens on serviceable delegator.
Signed-off-by: Wei Liu <wei.liu@zilliz.com>
issue: #37115
casue init query node client is too heavy, so we remove
updateShardClient from leader mutex, which cause much more concurrent
cornor cases.
This PR delay query node client's init operation until `getClient` is
called, then use leader mutex to protect updating shard client progress
to avoid concurrent issues.
---------
Signed-off-by: Wei Liu <wei.liu@zilliz.com>
Related to #35303
`deleteMut` shall be protecting streaming delete buffer, forward l0
could be move out of the rlock section to reduce tsafe impact from
loading segments.
Signed-off-by: Congqi Xia <congqi.xia@zilliz.com>
issue: #34298
because all vector index config checker has been moved into
vector_index_checker, then the useless checkers can be removed.
Signed-off-by: xianliang.li <xianliang.li@zilliz.com>
issue:https://github.com/milvus-io/milvus/issues/27576
# Main Goals
1. Create and describe collections with geospatial fields, enabling both
client and server to recognize and process geo fields.
2. Insert geospatial data as payload values in the insert binlog, and
print the values for verification.
3. Load segments containing geospatial data into memory.
4. Ensure query outputs can display geospatial data.
5. Support filtering on GIS functions for geospatial columns.
# Solution
1. **Add Type**: Modify the Milvus core by adding a Geospatial type in
both the C++ and Go code layers, defining the Geospatial data structure
and the corresponding interfaces.
2. **Dependency Libraries**: Introduce necessary geospatial data
processing libraries. In the C++ source code, use Conan package
management to include the GDAL library. In the Go source code, add the
go-geom library to the go.mod file.
3. **Protocol Interface**: Revise the Milvus protocol to provide
mechanisms for Geospatial message serialization and deserialization.
4. **Data Pipeline**: Facilitate interaction between the client and
proxy using the WKT format for geospatial data. The proxy will convert
all data into WKB format for downstream processing, providing column
data interfaces, segment encapsulation, segment loading, payload
writing, and cache block management.
5. **Query Operators**: Implement simple display and support for filter
queries. Initially, focus on filtering based on spatial relationships
for a single column of geospatial literal values, providing parsing and
execution for query expressions.
6. **Client Modification**: Enable the client to handle user input for
geospatial data and facilitate end-to-end testing.Check the modification
in pymilvus.
---------
Signed-off-by: tasty-gumi <1021989072@qq.com>
issue: #36672
The expression supports filling elements through templates, which helps
to reduce the overhead of parsing the elements.
---------
Signed-off-by: Cai Zhang <cai.zhang@zilliz.com>
issue: #37156
1. Still need to record the current stats version.
2. Set it to 0 when the current stats version is not found.
Signed-off-by: Cai Zhang <cai.zhang@zilliz.com>
Related to #36887
Remove non-hit pk delete record logic does not work since
`insert_record_.contain` does not work due to logic problem.
Signed-off-by: Congqi Xia <congqi.xia@zilliz.com>
See also #37205
Previously releasing growing segments could be triggered by two
conditions:
- Sealed Segment with same id is loaded
- Segment start position is before target checkpoint ts
Which has a worst case that the corresponding sealed segment is
compacted and the checkpoint is pinned by a growing l0 segment.
This PR introduces a new rule that: a growing segment could be released
if the segment id appeared in current target dropped segment id list.
Signed-off-by: Congqi Xia <congqi.xia@zilliz.com>
Related to #35303#30404
This PR change return type of `DeleteCodec.Deserialize` from
`storage.DeleteData` to `DeltaData`, which
reduces the memory usage of interface header.
Also refine `storage.DeltaData` methods to make it easier to usage.
Signed-off-by: Congqi Xia <congqi.xia@zilliz.com>
Related to #36102
Previous PR #36107 add grpc inteceptor to observe rpc stats. Using same
strategy, this pr add gin middleware to observer restful v2 rpc stats.
Signed-off-by: Congqi Xia <congqi.xia@zilliz.com>
Related to #36887
`LoadDeltaLogs` API did not check memory usage. When system is under
high delete load pressure, this could result into OOM quit.
This PR add resource check for `LoadDeltaLogs` actions and separate
internal deltalog loading function with public one.
---------
Signed-off-by: Congqi Xia <congqi.xia@zilliz.com>
Related to #36887
Previously using newly create pool per request shall cause goroutine
leakage. This PR change this behavior by using singleton delete pool.
This change could also provide better concurrency control over delete
memory usage.
Signed-off-by: Congqi Xia <congqi.xia@zilliz.com>
Also remove conflit check when executing L0. The exclusive is already
guarenteed in scheduler
See also: #37140
---------
Signed-off-by: yangxuan <xuan.yang@zilliz.com>
Related to #37183
Utilize proxy metacache for `HasCollection` request, if collection
exists in metacache, it could be deducted that collection must exist in
system.
Signed-off-by: Congqi Xia <congqi.xia@zilliz.com>
issue: #37158
Return the GuaranteeTS so that the subsequent requests following the
correct TS.
BeginTS is the current timestamp when the task is created.
The GuaranteeTS is the one parsed based on both consistency level and
beginTS, in PreExecute of the task on Proxy.
The delegator will wait until GuaranteeTS is met.
In PostExecute of the task on Proxy, the TS of the first iterator
request will be returned to the SDK and add it to the subsequent
requests.
Hence, if the default consistency level is Eventually or Bounded, the
order of TS will be
> Guarantee TS < BeginTS
If it returns the BeginTS, the second request will need to catch up and
result in extra 200ms max of latency, which results in something like
| Call | Latency |
| --- | --- |
| first call on `Next()` | 30ms |
| second call on `Next()` | 210ms |
| third call on `Next()` | 10ms |
| fourth call on `Next()` | 11 ms |
| ... | ... |
where we expect
| Call | Latency |
| --- | --- |
| first call on `Next()` | 30ms |
| second call on `Next()` | 10ms |
| third call on `Next()` | 10ms |
| fourth call on `Next()` | 11 ms |
| ... | ... |
Signed-off-by: Patrick Weizhi Xu <weizhi.xu@zilliz.com>
Related to #37177
Previous PR #37160
Collection meta is not ref-ed when loading l0 segment in `RemoteLoad`
policy, which cause collection meta release when lots of l0 segment
released.
Signed-off-by: Congqi Xia <congqi.xia@zilliz.com>
Timeout is a bad design for long running tasks, especially using a
static timeout config. We should monitor execution progress and fail the
task if the progress has been stale for a long time.
This pr is a small patch to stop DC from marking compaction tasks
timeout, while still waiting for DN to finish. The design is
self-conflicted. After this pr, mix and L0 compaction are no longer
controlled by DC timeout, but clustering is still under timeout control.
The compaction queue capacity grows larger for priority calc, hence
timeout compactions appears more often, and when timeout, the queuing
tasks will be timeout too, no compaction will success after.
See also: #37108, #37015
---------
Signed-off-by: yangxuan <xuan.yang@zilliz.com>
Related to #35303
Slice of `storage.PrimaryKey` will have extra interface cost for each
element, which may cause notable memory usage when delta row count
number is large.
This PR replaces PrimaryKey slice with PrimaryKeys interface saving the
extra interface cost.
---------
Signed-off-by: Congqi Xia <congqi.xia@zilliz.com>
issue: #36621
1. Add API to access task runtime metrics, including:
- build index task
- compaction task
- import task
- balance (including load/release of segments/channels and some leader
tasks on querycoord)
- sync task
2. Add a debug model to the webpage by using debug=true or debug=false
in the URL query parameters to enable or disable debug mode.
Signed-off-by: jaime <yun.zhang@zilliz.com>
Related to #37112
Skip load logic used to work only when there is multiple segment load
info entires in load request. In continous delete case, delegator still
loads l0 segment, which occupies lot of memory.
Signed-off-by: Congqi Xia <congqi.xia@zilliz.com>
OSPP 2024 project:
https://summer-ospp.ac.cn/org/prodetail/247410235?list=org&navpage=org
Solutions:
- parser (planparserv2)
- add CallExpr in planparserv2/Plan.g4
- update parser_visitor and show_visitor
- grpc protobuf
- add CallExpr in plan.proto
- execution (`core/src/exec`)
- add `CallExpr` `ValueExpr` and `ColumnExpr` (both logical and
physical) for function call and function parameters
- function factory (`core/src/exec/expression/function`)
- create a global hashmap when starting milvus (see server.go)
- the global hashmap stores function signatures and their function
pointers, the CallExpr in execution engine can get the function pointer
by function signature.
- custom functions
- empty(string)
- starts_with(string, string)
- add cpp/go unittests and E2E tests
closes: #36559
Signed-off-by: Yinzuo Jiang <jiangyinzuo@foxmail.com>
issue: https://github.com/milvus-io/milvus/issues/37083
We use vector of string_view to save data temporally but real string
data will be released after record batch is deconstructed.
Change it to vector of string to avoid memory corruption.
---------
Signed-off-by: sunby <sunbingyi1992@gmail.com>
Related to #35303
Delta data is not needed when using `RemoteLoad` l0 forward policy. By
skipping load delta data, memory pressure could be eased if l0 segment
size/number is large.
Signed-off-by: Congqi Xia <congqi.xia@zilliz.com>
issue: #37054
after querycoord restart, segment_checker may release segment by mistake
due to next target isn't ready yet.
This PR requires release segment must happens after next target is
ready.
Signed-off-by: Wei Liu <wei.liu@zilliz.com>
Related to #35303
This PR add metrics for querynode delegator delete buffer information,
which is related to dml quota logic.
Signed-off-by: Congqi Xia <congqi.xia@zilliz.com>
Relatedt #36887
DirectFoward streaming delete will cause memory usage explode if the
segments number was large. This PR add batching delete API and using it
for direct forward implementation.
Signed-off-by: Congqi Xia <congqi.xia@zilliz.com>
issue: #33550
cause wrong impl of UpdateCollectionNextTarget, if ReleaseCollection and
UpdateCollectionNextTarget happens at same time, the the released
partition's segment list may be add to target again, and delegator will
be marked as unserviceable due to lack of segment.
This PR fix the impl of UpdateCollectionNextTarget
Signed-off-by: Wei Liu <wei.liu@zilliz.com>
issue: #35576
This pr is to cover those cases when queryHook optimize search params
and make the result size insufficient, add retry search mechanism and
add related metrics for alarming.
---------
Signed-off-by: chasingegg <chao.gao@zilliz.com>
issue: #36686
This pr will remove pre-marking segments as L2 during clustering
compaction in version 2.5, and ensure compatibility with version 2.4.
The core of this change is to **ensure that the many-to-many lineage
derivation logic is correct, making sure that both the parent and child
cannot simultaneously exist in the target segment view.**
feature:
- Clustering compaction no longer marks the input segments as L2.
- Add a new field `is_invisible` to `segmentInfo`, and mark segments
that have completed clustering but have not yet built indexes as
`is_invisible` to prevent them from being loaded prematurely."
- Do not mark the input segment as `Dropped` before the clustering
compaction is completed.
- After compaction fails, only the result segment needs to be marked as
Dropped.
compatibility:
- If the upgraded task has not failed, there are no compatibility
issues.
- If the status after the upgrade is `MetaSaved`, then skip the stats
task based on whether TmpSegments is empty.
- If the failure occurs before `MetaSaved`:
- there are no ResultSegments, and InputSegments have not been marked as
dropped yet.
- the level of input segments need to revert to LastLevel
- If the failure occurs after `MetaSaved`:
- ResultSegments have already been generated, and InputSegments have
been marked as Dropped. At this point, simply make the ResultSegments
visible.
- the level of ResultSegments needs to be set to L1（in order to
participate in mixCompaction）
---------
Signed-off-by: Cai Zhang <cai.zhang@zilliz.com>
issue #34117
* Refactoring
* Added a capability to perform multiple bitwise `and` and `or`
operations in a single op
* AVX2, AVX512, ARM NEON, ARM SVE backed bitwise `and`, `op`, `xor` and
`sub` ops
* more unit tests for bitset
* fixed a bug in `or_with_count` for certain bitset sizes
* fixed a bug for certain offset values for inplace operations that take
two bitsets
Signed-off-by: Alexandr Guzhva <alexanderguzhva@gmail.com>
issue: #36686
bug reason:
- The clustering compaction tasks on the datanode were never cleaned up.
- The clustering compaction task contains a mapping from clustering key
to buffer, this caused a large memory leak.
fix:
- clean the tasks on datanode by datacoord when clustering compaction
finished.
- reset the mapping that from clustering key to buffer on datanode when
clustering finished.
Signed-off-by: Cai Zhang <cai.zhang@zilliz.com>
issue: #34553
when rootcoord trigger graceful stop progress, it will block until all
rpc finished. for create collection request, rootcoord need to block
until datacoord finish to watch all channels, but datacoord need to call
`rootcoord.Alloc` during watch channel, and rootcoord doesn't respond to
new request anymore. which cause create collection stucks, and graceful
stop progress stucks.
This PR remove the func call `rootcoord.Alloc` to solve the logic dead
lock during graceful stop progress.
Signed-off-by: Wei Liu <wei.liu@zilliz.com>
issue: #36868
if datacoord is syncing segments to datanode, and stop datacoord
happens, datacoord's stop progress will stuck until syncing segment
finished.
This PR add ctx to syncing segment, which will failed if stopping
datacoord happens.
Signed-off-by: Wei Liu <wei.liu@zilliz.com>
issue: https://github.com/milvus-io/milvus/issues/36835
currently searching BM25 output field using IP will end up in an error
in segcore which is hard to understand. now returning error in query
node delegator and provide more useful error message
Signed-off-by: Buqian Zheng <zhengbuqian@gmail.com>
Related to #36887
Forward delete to L0 segment will return error and mark l0 segment
offline causing delegator unserviceable
Signed-off-by: Congqi Xia <congqi.xia@zilliz.com>
Start position & level info is missing for growing segment loaded in
watch dml channel operation.
Level is important for metrics and start position is crucial for growing
exclude logic.
Signed-off-by: Congqi Xia <congqi.xia@zilliz.com>
In quota center, ignore the "DB not found error" to prevent it from
affecting the rate limiting of other databases.
/kind improvement
---------
Signed-off-by: bigsheeper <yihao.dai@zilliz.com>
wei liu
XuanYang-cn
cai.zhang
Zhen Ye
Buqian Zheng
congqixia
nish112022
yihao.dai
SimFG
cqy123456
Xiaofan
Bingyi Sun
smellthemoon
MimoN
sthuang
aoiasd
jaime
Chun Han
foxspy
zhagnlu
MimoN
Ted Xu
wayblink
Yinzuo Jiang
zhenshan.cao
Hao Tan
liliu-z
Gao
Patrick Weizhi Xu
Yinzuo Jiang
yellow-shine
Alexander Guzhva
foxspy
OxalisCu