Describe the feature:
Elasticsearch version (
bin/elasticsearch --version): 6.2.2
Plugins installed: [no plugin]
JVM version (
java -version): 1.8.0
OS version (
uname -a if on a Unix-like system): inux tjwq01-ibu-video002158.tjwq01.ksyun.com 2.6.32-431.el6.x86_64 #1 SMP Fri Nov 22 03:15:09 UTC 2013 x86_64
500g mem. Total. 2 es (2.1.2) use 31 * 2 g mem
email@example.com ~/es6_2/node1/elasticsearch-6.2.2]$ free -g │ 27.43 0.00 13.06 0.75 0.00 58.76 total used free shared buffers cached │ Mem: 504 121 382 0 0 35
Description of the problem including expected versus actual behavior: When start es. There will be one core to locate memory. My machine had 2 process already running (es 2.1.2 ) . When I ran the third es (6.2.2) . The es process allocate memroy very slowly ( when set xms = 21g, It cost 30min).
When I comment jvm.options:alwayspretouch
line. It also cannot work well
Steps to reproduce:
Please include a minimal but complete recreation of the problem, including
(e.g.) index creation, mappings, settings, query etc. The easier you make for
us to reproduce it, the more likely that somebody will take the time to look at it.
- hard to produce. I have 5 es machines. 3 allocating memory is fast. these 2 are slow. all the 5 nodes have 2 es already ran on them as data nodes(es 2.1.2)
Provide logs (if relevant):
when try jstack es62 process. the es62 will core dump.