2 Couchbase type buckets slower than 1
Hi guys,
I'm doing some stress testing with couchbase. I've 2 windows machines each with 8GB RAM and 2 cores.
So, i've a .net client sending fixed size packages of 100KB until it reaches a total of 8GB. It's running parallel tasks doing 70% of GETs and 30% of PUTs.
First, i was using just one of this two servers and i had 5,4 minutes to complete this 8GB information with Avg operation time of 19ms-GET and 33ms-PUT. Then, with 2 servers, with a total of 8GB RAM for the bucket, it took 7 minutes to finish the 8GB info, with avg op time of 53ms-GET and 89ms-PUT.
Can you help me? is there some config settings i should do?
Another point is the blank spaces i regularly get in the operations monitor. It looks like the bucket is freezing. I can post some print screen if you need. Any ideas?
This is a vital point in my structure.
Thanks a lot
Hi,
Are they dedicated to couchbase. otherways any other application on these nodes running apart from couchbase.
Could you please check the logs. Also let me know the network details. Hope they are physical machines. Screenshots might help.
Hi Neo,
Yes, they are dedicated to couchbase. Both machines are into the cloud, not physical ones. I'll run another tests so i can get some screenshots for you.
Thanks
Hi,
Could you please also add some more details about the Bucket, RAM allocated per bucket and replicas info.
Thanks
Neo
I wouldn't expect Couchbase to take any more than 1ms for an get or set request. Can you post the code that you wrote that causes you to see these high latencies?