JSON load balancer config file #328
Conversation
|
@andreaseno I merged #327 so now you will need to resolve conflicts. |
| @@ -600,6 +674,10 @@ main(int argc, char *argv[]) { | |||
| int arg_offset; | |||
| const char *progname = argv[0]; | |||
|
|
|||
| time_t t; | |||
| /* Intializes RANDOM number generator */ | |||
| srand((unsigned) time(&t)); | |||
There was a problem hiding this comment.
What is the issue with the indentation here?
There was a problem hiding this comment.
looks like tabs not spaces. @catherinemeadows -- do you know if we have a working linter script that checks for this kind of formatting issue in ONVM? I thought our GitHub actions would do it automatically, but it seems not.
There was a problem hiding this comment.
Yes, in /openNetVM/style, there is documentation about manually running our linter scripts. I think the GitHub actions weren't triggered yet because there are still merge conflicts.
examples/load_balancer/server.json
Outdated
| { | ||
| "Info" : { | ||
| "list_size": 2, | ||
| "policy": "random" |
| @@ -555,6 +628,7 @@ packet_handler(struct rte_mbuf *pkt, struct onvm_pkt_meta *meta, | |||
| for (i = 0; i < RTE_ETHER_ADDR_LEN; i++) { | |||
| flow_info->s_addr_bytes[i] = ehdr->s_addr.addr_bytes[i]; | |||
| } | |||
| if(debug_mode) printf("New connection made with server %d.\n",flow_info->dest); | |||
There was a problem hiding this comment.
Can we have this display the src IP and port? Something like:
192.168.1.100:4256 assigned to server 2
|
@andreaseno @xwedea I added a couple more changes. If you can fix them sometime in the next few weeks - great! If not I'll get another student to help out later. |
Replaced the server.conf file inside the load balancer nf with a json file called server.json. Because the config file parser was written for a plaintext file rather than a json file, I had to rewrite the backend config file parse function. It now uses cjson to parse the config file.
Summary:
Usage:
Merging notes:
TODO before merging :
Test Plan:
I tested the changes by doing a full setup of a 2 server ONVM experiment, made sure the arp_response and load balancer nf's ran correctly without error, and used iperf to verify that the packets would still send correctly. The only other necessary testing I can think of off the top of my head is verifying it works with more than 2 servers.
Review:
(optional) << @-mention people who should review these changes >>
(optional) Subscribers: << @-mention people who probably care about these changes >>