Huge virtual memory usage

[excieve]

I've been tracing resource usage with my benchmarks and noticed that couch-chakra processes are using about 35G of virtual memory (with only ≈55M RSS):

#      Time   UID       PID    %usr %system  %guest    %CPU   CPU  minflt/s  majflt/s     VSZ     RSS   %MEM   kB_rd/s   kB_wr/s kB_ccwr/s iodelay threads   fd-nr  Command
 1489004421  1000     13690  100,00   65,02    0,00   65,30     6 193037,44      0,00 2980360  799660   4,89      0,00   1442,36      0,00       0      29      77  /opt/couchdb/bin/../erts-6.2/bin/beam.smp -K true -A 16 -Bd -- -root /opt/couchdb/bin/.. -progname couchdb -- -home /opt/couchd
 1489004421  1000     14130    0,49    0,00    0,00    0,06     6      0,00      0,00  228460   25420   0,16      0,00      0,00      0,00       0       1       4  python3 utils/addview.py data/reference/income_map_function.es6 -u admin -p admin -D map_tests_chakra -v read_income_test -l ch
 1489004421  1000     14134   24,63    0,99    0,00    3,23     1    120,20      0,00 35836712   56364   0,34      0,00      0,00      0,00       0       7       9  /opt/couch-chakra/couch-chakra /opt/couch-chakra/main.js 
 1489004421  1000     14143   13,79    0,00    0,00    1,74     2     15,76      0,00 35834916   54580   0,33      0,00      0,00      0,00       0       7       9  /opt/couch-chakra/couch-chakra /opt/couch-chakra/main.js 
 1489004421  1000     14144   22,66    0,49    0,00    2,92     6     67,00      0,00 35835356   56140   0,34      0,00      0,00      0,00       0       7       9  /opt/couch-chakra/couch-chakra /opt/couch-chakra/main.js 
 1489004421  1000     14145   22,17    0,49    0,00    2,86     4     83,25      0,00 35834736   55340   0,34      0,00      0,00      0,00       0       7       9  /opt/couch-chakra/couch-chakra /opt/couch-chakra/main.js 
 1489004421  1000     14146   20,20    0,49    0,00    2,61     1     86,70      0,00 35835912   55884   0,34      0,00      0,00      0,00       0       7       9  /opt/couch-chakra/couch-chakra /opt/couch-chakra/main.js 
 1489004421  1000     14147   26,11    0,49    0,00    3,35     2    127,59      0,00 35835208   55868   0,34      0,00      0,00      0,00       0       7       9  /opt/couch-chakra/couch-chakra /opt/couch-chakra/main.js 
 1489004421  1000     14148   20,69    0,49    0,00    2,67     7    183,74      0,00 35769912   55496   0,34      0,00      0,00      0,00       0       7       9  /opt/couch-chakra/couch-chakra /opt/couch-chakra/main.js 
 1489004421  1000     14149   23,15    0,49    0,00    2,98     5     22,17      0,00 35761756   49764   0,30      0,00      0,00      0,00       0       7       9  /opt/couch-chakra/couch-chakra /opt/couch-chakra/main.js

I'm pretty sure it shouldn't be like this. Other query servers (and even CouchDB itself) don't use anywhere near this much. This is maintained even when no data is being processed.

[dmunch]

What a hungry little beast, that's unexpected! I'm not entirely sure if that comes from the way I use ChakraCore or if it's the runtime itself which just behaves this way.

One thing you could try is to use the ´-S´ switch as a parameter to chakra-core on the command line, which will call JsSetRuntimeMemoryLimit.

[excieve]

-S 200000000 didn't help with this, still ≈35GB of virtual mem. I think it only affects real memory.