Replies: 1 comment 2 replies
|
TSocket read 0 bytes means connection was closed by server unexpectedly. common causes:
# increase timeout
config = Config()
config.timeout = 60000 # ms
pool.init([...], config)
# increase pool size
session_pool = SessionPool(
user_name="root",
password="nebula",
space_name="your_space",
addresses=[("127.0.0.1", 9669)],
pool_size=30 # increase from default
)
from tenacity import retry, stop_after_attempt
@retry(stop=stop_after_attempt(3))
def execute_query(session_pool, query):
return session_pool.execute(query)
docker logs nebula-graphdlook for OOM or connection errors
what version of nebula are you using? and is this happening under high load? |
|
Hi, Thanks for the reply I am using version 3.8 It is happening on an instance where requests are either bursts once midnight and then sparse through the day, and this errors seems to be happening mostly when the traffic is very sparse i have checked the memory limits and that didn't seem to be the issue and also i haven't seen this error log on the instances which runs whole day, so connection pool exhaustion and server resource limit issue didn't seem possible for my case I checked the logs of the graphd and there were logs of session expiration and when i learned when unset the session never expires on client but server expires them every 8 hr which might cause expired session reference: https://discuss.nebula-graph.com.cn/t/topic/12274/4?u=yochi for server timeout # nebula3/Config.py
class Config(object):
# the min connection always in pool
min_connection_pool_size = 0
# the max connection in pool
max_connection_pool_size = 10
# connection or execute timeout, unit ms, 0 means no timeout
timeout = 0
# 0 means will never close the idle connection, unit ms,
idle_time = 0
# the interval to check idle time connection, unit second, -1 means no check
interval_check = -1
# use http2 or not
use_http2 = False
# headers for http2, dict type
http_headers = NoneSo i set the config timeout to be 10 min and interval check to 1 min on the client but that also didn't seem to fix this issue I am also thinking of adding the retry logic but i am not sure if it might lead to double creation of nodes or edges in the Insert statements, let me know if i am wrong here i found this post also mentioning the same issue: https://discuss.nebula-graph.com.cn/t/topic/12579 Please let me know if there is any better solution or any other reason that might have caused the issue based on the above information |
|
for reference these were the logs in nebula-grapg.INFO |
Uh oh!
There was an error while loading. Please reload this page.
hi everyone, i am randomly getting these errors
while running INSERT VERTEX command or LOOKUP commands
i am also not able to find these error code on https://docs.nebula-graph.io/3.8.0/20.appendix/error-code/
for any reference
can anyone please help for any possible reason for this to be happening and how this can be solved
All reactions