In one of our SPLUNK SH Cluster environment, inserting records into KVStore using rest command is taking lot of time
and I am able to post only 20 Records into the collection in one minute.
The splunkd_access.log shows below message and tells that the each KVStore post is taking close to 3-5 secs.
splunkd_access.log output as below.
127.0.0.1 - user [06/Jan/2020:13:56:55.591 +1100] "POST /servicesNS/nobody/test_app/storage/collections/data/TestCollection HTTP/1.1" 201 13 - - - 5889ms
There is nothing suspicious message in mongo logs which tells why the insert is taking lot of time.
Here is the sample script I am using for inserting records into KVStore. I can use the outputlookup command to update records into KVStore but this is to investigate the issue which we are facing in overall slowness of KVStore.
import json
import requests
headers = {
'Content-Type': 'application/json',
}
with open('test.json') as json_file:
data = json.load(json_file)
for p in data['definitions']:
print p;
response = requests.post('https://localhost:8089/servicesNS/nobody/test_app/storage/collections/data/TestCollection', headers=headers, verify=False, auth=('user', 'password'),data=json.dumps(p))
I am raising with splunk support in parallel posting here and would like to check if anyone have had these kind of issues in their environment and what steps were taken to figure out what is causing this issue.
Storage IO wait on the mount point is close to 20 ms where SPLUNK is installed.
We were using an LDAP user for authentication and the authentication was taking more time , hence it was slowing the KVStore update using API.
We were using an LDAP user for authentication and the authentication was taking more time , hence it was slowing the KVStore update using API.