Hi Team, I am using event store in Docker on Ubuntu 20.04 Server on Lenovo ST 500 Machine. I am getting delayed while persisting event to event store but when I switched the same to my laptop then it’s working fine and taking very less time.
When on Ubuntu Server Response Time : 300 to 400 ms (Only EventStore Operation)
When on Lenovo i5 Laptop: 20ms (Only EventStore Operation)
Also, same code is working fine on Dell PowerEdge Series with all same configuration !!
Well Docker is a layer of virtualization that adds a virtual filesystem and network on top of the guest OS filesystem and network.
But how same code is working fine on other machine with same configuration and environment. Is it issue of OS or Hardware or Docker and how to overcome this issue?
Apparently, Lenovo is your machine, and you don’t have any networking timeout. The container performance heavily depends on the underlying machine performance, plus the specifics of the network and disk drivers used.
You can try to find out where the issue is:
Hi @alexey.zimarev, Thanks for your suggession !!
Can you let us know, how to implement ESDB in an In-Memory store in docker on ubuntu 20.04?
Also, Kafka & MongoDb is already running on same server with nice performance.
Can you let us know, how to implement ESDB in an In-Mem …
The command line swtich is
-MemDb=true
if using env variable
EVENTSTORE_MEMDB=true
Also, Kafka & MongoDb is already running on same server with nice performance.
The question was : do you see any perfomance difference on the different machine you are running them on ?
Hi @yves.lorphelin, Thanks for your quick response !!
I am using EventStore suggested docker-compose as below, can you suggest where to modify in docker-compose file.
For point 2, I will check the same and let you know in sometimes !!
Yves already wrote it.
Hi @alexey.zimarev, Thanks for your quick response !!
Allow me sometime to test and let you know with outcomes but also for your information that the same eventstore in working fine on Ubuntu 22.04 inside Docker of Dell PowerEdge R340 and response time is terrific !!
Kafka can deliver high message publishing throughput when you use writes and delayed acknowledgement. If you explicitly wait for each message to be acknowledged, you’ll see reduced throughput, and it will tell you the real latency per message.
MongoDB can also be configured to write to the cache and acknowledge. It’s not reliable, and it’s no longer the default setting for many years, but without knowing your configuration I can’t be sure.
Hi @alexey.zimarev & @yves.lorphelin, Thank you so much for wonderful suggestion. It’s working faster than ever. Is there any issue to continue with an In-Memory Db?
Yes, nothing gets saved to the disk, as the option implies. When you restart the container, everything will be lost.
