During the convening in Zanzibar in late 2022 , I brought along a 4GB Intel NUC computer and demonstrated mini-loop deploying Mojaloop v14.x (probably release candidate). It was a bit slow on the NUC as it was a bit tight on memory but the helm deploy of the Mojaloop chart succeeded and subsequently the testing toolkit could be executed just fine. Back then we were trying to keep the memory use low for demo/test so as to enable (I thought) as many folks as possible to run on their laptops as I had assumed that most folks in East Africa would have 4GB laptops.
With that as background , I have been working on updating mini-loop to deploy and test the recent Mojaloop v15.0.0 release and also adding some instrumentation and extra checks to ensure the prerequisites for running Mojaloop are met. I am working on a 16GB VM (In Azure) and I noticed that when I deploy Mojaloop plus 3PPI and Bulk payments that 16 GB is barely enough and some of the pods seem to get get pre-empted by kubernetes during deployment. So I added some more instrumentation and then ran with 3PPI and bulk disabled to compare against that earlier experience with Mojaloop v14.1 (without yet going back and actually running v14.1)
I tested against k8s v1.25 and v1.26 (earlier 1.24) and all used about the same memory and I used mini-loop v5.0 (note: it is work in progress) but you can check it out and try it
git clone --branch v5.0 https://github.com/tdaly61/mini-loop.git
I haven't had time to investigate to thoroughly yet but below is the data reported from the latest versions of mini-loop deploy tool and I would be interested in any one else's observations or experience with running v15.0.0
********* mini-loop stats *******************************
kubernetes distro:version [k3s]:[v1.26]
installation options []
Number of pods running [46]
major processing times :
helm_install_ml: 718 seconds
helm_install_be: 263 seconds
repackage_ml: 162 seconds
install_ml: 728 seconds
[ Oh and note: if it appears like there is missing RAM, it is used by the page cache in Linux , so the most important column here is "RAM used" ]
Total system memory: 15.6284 GB
When | RAM used | RAM free | RAM used %
at_start | 1.11GB | 11.22GB | 9.00%
at_end | 8.80GB | 0.80GB | 91.67%
************ mini-loop stats ******************************
********* mini-loop stats *******************************
kubernetes distro:version [microk8s]:[v1.26]
installation options []
Number of pods running [45]
major processing times :
helm_install_ml: 1273 seconds
helm_install_be: 312 seconds
repackage_ml: 168 seconds
install_ml: 1284 seconds
Total system memory: 15.6284 GB
When RAM used RAM free RAM used %
at_start | 1.18GB | 10.62GB | 10.03%
at_end | 8.81GB | 0.77GB | 92.00%
Top comments (0)