HuangWei

comment about the time units of zk

1

e.g. https://github.com/4paradigm/OpenMLDB/blob/79b34500d16c56cc68d5defb3f17092b08fa41b1/src/flags.cc#L21 https://github.com/4paradigm/OpenMLDB/blob/79b34500d16c56cc68d5defb3f17092b08fa41b1/src/flags.cc#L28 We should add the time unit, all milliseconds. And add it in user doc too.

log: if CLI remains open for a long time, get so many sampler warn log

3

``` W0712 10:15:22.106148 129511424 sampler.cpp:189] bvar is busy at sampling for 2 seconds! ``` brpc log, maybe we should avoid to print it on terminal too, redirect to log file.

**Bug Description** insert a large number( > int64 max 2147483647) into timestamp column, won't get error. Although, it won't add one row. **Expected Behavior** get error msg? **Steps to Reproduce**...

Each apiserver request has its own db, but NsClient used to cache db. So we may get no db bug when we request apiserver and it calls NsClient.

glog config can't use gflags style to config now, cuz we built glog before gflags in thirdparty pre-built. e.g. `--v`, `--vmodule` Now we can use `GLOG_v` to open vlog. But...

ut: alloc-dealloc-mismatch (operator new [] vs free) in api_server_test

1

Ref #358 ``` ==7595==ERROR: AddressSanitizer: alloc-dealloc-mismatch (operator new [] vs free) on 0x60600024df40 ``` https://github.com/4paradigm/OpenMLDB/blob/326295840f6c2932c2d47937f8cf0ddf6c238664/hybridse/include/base/fe_slice.h#L104 is new [], we should use delete [], not free.

Cpp ut defines zk port 6181 in code. If we want to modify it, we should rebuild the binary files. It's not good for testing. We should support: ``` ./build/bin/xxx_test...

ut: api server test coredump in teardown on macOS

1

``` * frame #0: 0x00007fff6a79304a libsystem_kernel.dylib`madvise + 10 frame #1: 0x00007fff6a81093e libsystem_malloc.dylib`mvm_madvise_free + 87 frame #2: 0x00007fff6a809d9e libsystem_malloc.dylib`free_small + 1464 frame #3: 0x0000000103bcb407 api_server_test`llvm::orc::AbsoluteSymbolsMaterializationUnit::~AbsoluteSymbolsMaterializationUnit() + 87 frame #4: 0x0000000103be8e65 api_server_test`llvm::orc::JITDylib::~JITDylib()...

https://github.com/4paradigm/OpenMLDB/blob/053c4e3d885b237711d3719be7bf0c1905c5197e/java/openmldb-spark-connector/pom.xml#L37 0.3.0 too old. And we should use the project version. Profile is needless

batch: bug in handle key if is null

2

https://github.com/4paradigm/OpenMLDB/blob/908e9edbe05a86dccdbf8d1579bad88454480d59/java/openmldb-batch/src/main/scala/com/_4paradigm/openmldb/batch/nodes/WindowAggPlan.scala#L424-L425 We get key by `extractKey`, and then check it by `isValidOrder`. After #1408 fixed, `isValidOrder` arg is java Long, it may be null or a long. But `extractKey` returns...