1. add coordinator graceful stop timeout to 5s
2. change the order of datacoord component while stop
3. change querynode grace stop timeout to 900s, and we should
potentially change this to 600s when graceful stop is smooth
issue: #30310
also see pr: #30306
---------
Signed-off-by: chyezh <chyezh@outlook.com>
Compaction would copy logPaths from comapctFrom segA to compactTo segB,
and previous code would copy the logPath directly, causing there're
full-logPaths-of-segA in compactTo segB's meta. So, for the next
compaction of segB, if segA has been GCed, Download would report error
"The sperified key not found".
This PR makes sure compactTo segment's meta contains logID only. And
this PR also refines CompleteComapctionMutation, increasing some
readability and merge two methods into one.
See also: #30496
Signed-off-by: yangxuan <xuan.yang@zilliz.com>
Support get sdk type by user agent when we can't get sdk version by
connection in access log.
---------
Signed-off-by: aoiasd <zhicheng.yue@zilliz.com>
/kind improvement
this removes the 1x copying while loading variable length data, also
avoids constructing std::string, which could lead to memory
fragmentation
---------
Signed-off-by: yah01 <yah2er0ne@outlook.com>
Signed-off-by: longjiquan <jiquan.long@zilliz.com>
Co-authored-by: yah01 <yah2er0ne@outlook.com>
OpenShift runs Pods with a random uid and gid 0.
As Milvus needs to write into the /milvus directory, this fix modifies
the group permissions to allow the root group (gid 0) to write into it.
This is a simple solution that allows to run Milvus on OpenShift without
adding security constraints, use special ServiceAccounts or whatever.
This should also not impact any other kubernetes platform as it's a
simple group permission change, nothing regarding the userid.
Note: Other actions are necessary to properly deploy the full stack
(Minio, Pulsar,...) on OpenShift. I will document them in the helm chart
project.
Fixes#25565
Signed-off-by: Guillaume Moutier <guillaume.moutier@gmail.com>
issue: #30150
`checkLeaderTaskStale` will check segment whether exist on next current
for leaderTask's growing action, which will cause promote leader task
failed when segment only exist on current target
This PR will check segment for both current or next target.
Signed-off-by: Wei Liu <wei.liu@zilliz.com>
if check in Segcore, will not do the it when not insert data.
so, check "radius" and "range_filter" in proxy.
related with #30365
Signed-off-by: lixinguo <xinguo.li@zilliz.com>
Co-authored-by: lixinguo <xinguo.li@zilliz.com>
See also #30832
This PR removes time tick delay metrics when rootcoord GetMetrics
response does not have previously existed querynode/datanode
Also add unit tests for this case
---------
Signed-off-by: Congqi Xia <congqi.xia@zilliz.com>
Signed-off-by: Congqi.Xia <congqi.xia@zilliz.com>
bug list
1. element data type is needed while create a collection with an array
field #30638
2. spelling mistake about metricsType #30643
3. new collection enable dynamic field / auto id as user defined #30665
4. convert rowCount from string to int #30661
5. try it's best to create a new collection #30652
6. int64 percious #20415
7. insert into collection which has multi vector fields #30674
8. cannot rename a collection to other database #30700
9. update the request body of "indexes/create" #30769
10. got [] while list indexes of a collection which has no index #30722
11. restful need encode password before call
CreateCredential/UpdateCredential #30730
12. some parameters missed the required label #30737
13.define the field to be or not to be a partition key while create a
collection #30797
enhance: support opentelemetry
enhance: support dataType: Float16Vector & BFloat16Vector #22837
enhance: describe collection will show the field is partition key or not
#30789
Signed-off-by: PowderLi <min.li@zilliz.com>
See also #30756
This PR:
- Request disk resource when index type, version loaded with disk
- Add attribute cache for index utility
- Add `typeutil.Pair`
---------
Signed-off-by: Congqi Xia <congqi.xia@zilliz.com>
patch search cache param from index configs when index meta could not
get the search cache size key
#issue: #30113
Signed-off-by: xianliang.li <xianliang.li@zilliz.com>
issue: #30150
see also: #30258
cause `SyncDataDistribution` will try to load delta for segment. if miss
indexInfo in request, sync action will failed due to lack of index info.
This PR set indexinfo when try to set segment to leader view
Signed-off-by: Wei Liu <wei.liu@zilliz.com>
