Hi Prashant, Can you describe the steps you use to run `jeprof` in detail? In my case, I did it by logging in to Taskmanager's shell command line and then operating it through shell commands. But I am confused that I saw the curl operation in the error log you provided.
Also, it is true that rocksdb is not perfect in memory control and it is possible to exceed the managed memory limit, you can refer to the documentation[1] for more details. [1] Write buffer manager internals - Google 文档 <https://docs.google.com/document/d/1_4Brwy2Axzzqu7SJ4hLLl92hVCpeRlVEG-fj8KsTMUo/edit#heading=h.f5wfmsmpemd0> Best, Yu Chen > 2023年12月5日 04:42，prashant parbhane <parbhane....@gmail.com> 写道： > > Hi Yu, > > Thanks for your reply. > > When i run below script > > ``` > jeprof --show_bytes -svg `which java` /tmp/jeprof.out.301.1009.i1009.heap > > 1009.svg > ``` > i am getting below error > > ``` > Gathering CPU profile from http:///pprof/profile?seconds=30 for 30 seconds to > /root/jeprof/java.1701718686. > Be patient... > Failed to get profile: curl -s --fail --max-time 90 > 'http:///pprof/profile?seconds=30' > /root/jeprof/.tmp.java.1701718686.: No > such file or directory > ``` > Any input on this? > > However, oomkill was resolve with below rocksdb configurations > • "state.backend.rocksdb.memory.managed": "false", > "state.backend.rocksdb.block.cache-size": "10m", > "state.backend.rocksdb.writebuffer.size": "128m", > "state.backend.rocksdb.writebuffer.count": "134217728" > "state.backend.rocksdb.ttl.compaction.filter.enabled":"true" > > > Thanks, > Prashant > > On Mon, Nov 27, 2023 at 7:11 PM Xuyang <xyzhong...@163.com> wrote: > Hi, Prashant. > I think Yu Chen has given a professional troubleshooting ideas. Another thing > I want to ask is whether you use some > user defined function to store some objects? You can firstly dump the memory > and get more details to check for memory leaks. > > -- > Best！ > Xuyang > 在 2023-11-28 09:12:01，"Yu Chen" <yuchen.e...@gmail.com> 写道： > Hi Prashant, > > OOMkill was mostly caused by workset memory exceed the pod limit. > We have to first expand the OVERHEAD memory properly by the following params > to observe if the problem can be solved. > ``` > taskmanager.memory.jvm-overhead.max=1536m > taskmanager.memory.jvm-overhead.min=1536m > ``` > > And if the OOMKill still exists, we need to suspect if the task has an > off-heap memory leak. > One of the most popular tools, jemallc, is recommended. You have to install > the jemalloc in the image arrording to the document[1]. > After that, you can enable jemalloc profiling by setting environment for the > taskmanager: > ``` > containerized.taskmanager.env.MALLOC_CONF=prof:true,lg_prof_interval:30,lg_prof_sample:16,prof_prefix:/tmp/jeprof.out > ``` > After running for a while, you can log into the Taskmanager to generate svg > files to troubleshoot off-heap memory distribution. > ``` > jeprof --show_bytes -svg `which java` /tmp/jeprof.out.301.1009.i1009.heap > > 1009.svg > ``` > > Otherwise, if the OOMKill no longer occurs, but the GC overhead limit > exceeded, then you should dump heap memory to find out what objects are > taking up so much of the memory. > Here is the command for you. > ``` > jmap -dump:live,format=b,file=/tmp/heap.hprof <pid> > ``` > > [1] Using jemalloc to Optimize Memory Allocation — Sentieon Appnotes > 202308.01 documentation > > Best, > Yu Chen > 发件人: prashant parbhane <parbhane....@gmail.com> > 发送时间: 2023年11月28日 1:42 > 收件人: user@flink.apache.org <user@flink.apache.org> > 主题: oomkill issue Hello, > > We have been facing this oomkill issue, where task managers are getting > restarted with this error. > I am seeing memory consumption increasing in a linear manner, i have given > memory and CPU as high as possible but still facing the same issue. > > We are using rocksdb for the state backend, is there a way to find which > operator causing this issue? or find which operator takes more memory? Any > good practice that we can follow? We are using broadcast state. > > Thanks, > Prashant
