JinLiOnline opened a new issue #5794:
URL: https://github.com/apache/incubator-doris/issues/5794


   **Describe the bug**
   doris13版本,BE 发生oom, 
操作系统可用内存700G,top命令中显示,be只占用600G,其他进程占用内存很小。但是be却发生oom,被系统kill掉。
   
   经过排查发现,查看be的8040端口  
   $ curl http://127.0.0.1:8040/memz
   
   <pre>Mem Limit: 603.26 GB
   Mem Consumption: 29.70 GB
   
   MALLOC:   40138085404224 (38278661.2 MiB) Bytes in use by application
   MALLOC: +            0 (    0.0 MiB) Bytes in page heap freelist
   MALLOC: +  19226145608 (18335.5 MiB) Bytes in central cache freelist
   MALLOC: +       262144 (    0.2 MiB) Bytes in transfer cache freelist
   MALLOC: +    843947640 (  804.9 MiB) Bytes in thread cache freelists
   MALLOC: +  42186440704 (40232.1 MiB) Bytes in malloc metadata
   MALLOC:   ------------
   MALLOC: = 40200342200320 (38338033.9 MiB) Actual memory used (physical + 
swap)
   MALLOC: + 1512264359936 (1442207.7 MiB) Bytes released to OS (aka unmapped)
   MALLOC:   ------------
   MALLOC: = 41712606560256 (39780241.5 MiB) Virtual address space used
   MALLOC:
   MALLOC:       28652164              Spans in use
   MALLOC:           1010              Thread heaps in use
   MALLOC:           8192              Tcmalloc page size
   
   Call ReleaseFreeMemory() to release freelist memory to the OS (via 
madvise()).
   Bytes released to the OS take up virtual address space but no physical 
memory.
   
   
   上面显示malloc 38278661.2MB ==> 36T, 远超物理内存。
   
   查看pagetables发现:
   
   $ cat /proc/meminfo | grep PageTables
   PageTables:     79106020 kB
   
   换算得 75G。
   
   pagetables吃掉 75G,be占用600G,导致操作系统内存不足,触发oom干掉消耗内存最大的 be进程。
   
   


-- 
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.

For queries about this service, please contact Infrastructure at:
[email protected]



---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]

Reply via email to