Any known issues with circuit_breaking_execption?


#1

hi,
we have 2 clusters that share the same data . the difference between them is one with ROR and the other isn’t.
the cluster with the ROR ,one of it’s nodes , from time to time get the above error and from that moment it’s need to restart (since every command/query gets that error).

did you come across a problem like this?

thanks


(Simone Scarduzio) #2

Hi @sdba2, Can you paste the stack trace of the error?


#3

hi,
it’s difficult for me since this occur on an offline machine (not connect to the internet).
the message looks like (i typed it, hope i didn’t miss) :

{
“error”:{
“root_cause”:[
{
“type”:“circuit_breaking_exception”,
“reason”:"[parent] Data too large, data for [<http_request>] wold be[21578640180/20g], which is larger than the limit of [175584431334/16.3gb]",
“bytes_wanted”: 21578640180,
“bytes_limit”: 175584431334
}
],
“type”: “circuit_breaking_exception”,
“reason”:"[parent] Data too large, data for [<http_request>] wold be[21578640180/20g], which is larger than the limit of [175584431334/16.3gb]",
“bytes_wanted”: 21578640180,
“bytes_limit”: 175584431334
},
“status”: 503
}

thanks


(Simone Scarduzio) #4

In what version of ES does this happen? BTW I fixed a memory leak in the current version, after you tell me your ES version, I will give you a new build to see if you can reproduce.


#5

the es version is 5.4.0


(Simone Scarduzio) #6

Have you seen this fix? It’s been fixed in 6.0.x