Closed kevindweb closed 5 years ago
@onvm you there now?
@onvm you there now?
Your results will arrive shortly
@onvm you there now?
Error: ERROR: Script failed on nimbnode17
@onvm you know that was develop right?
@onvm you know that was develop right?
Your results will arrive shortly
To be clear, the first "failure" was not on the Pktgen branch of nn44 ci. Nimbnode17 got hung on setting up the environment. CI was running on the develop branch both times it posted here. See the latest post on #147 for real pktgen results.
@onvm check pktgen please
@onvm check pktgen please
Your results will arrive shortly
ci's got it! we have to figure out a good median for pktgen (maybe after merging the flow table macros...)
@kevindweb I've merged #143, update this pr so its up to date. Also add the CI stuff(disabling flow table lookup here) + update our benchmarks for speed tester (just so we're floating roughly at 100%, as its currently always at 109%)
@koolzz thanks for the comments! I'll work on updates for these this weekend. As for the file restructuring, I mentioned I need to change the helper-functions because for example, manager.sh
only really needs the run_linter
, fetch_files
, and print_header
functions. Worker.sh however doesn't need run_linter
or fetch_files
, but needs print_header
, build_onvm
, and install_env
. The only commonality is printing. For this reason, I'll make it more concise and possibly rename the file. It just made no sense to have a global helper function with no common functions. So I made a worker folder, which helps with scp (it's cleaner with all these new files), as well as organized a cluttered ci folder.
*edit, check_exit_code
is also used, but should be different, because a worker node does not have access to post-msg.py
, which causes errors occasionally.
@onvm test the changes, please
@onvm test the changes, please
Your results will arrive shortly
I have to get a handle on how sporadic these nimbnode results are. @koolzz any suggestions? I just received when tested 40mil+, how could it be such a big difference?
Are you using a different node to get these speed tester results? We were rather stable with speed tester tests before
These results are both from nimbnode17. I figured it made more sense to run pktgen and speed_test on the same node during a single CI run, is that not a good plan? I will look into why some runs are fast than others, though. Just weird because some runs are so much better than normal ~41 mil I've seen
Testing
Your results will arrive shortly
Testing
Your results will arrive shortly
Testing
Your results will arrive shortly
Testing
Your results will arrive shortly
@onvm let's see develop completely merged
@onvm let's see develop completely merged
Your results will arrive shortly
@onvm it's difficult to run Pktgen when the link is down right?
@onvm it's difficult to run Pktgen when the link is down right?
Your results will arrive shortly
@onvm can we try nn30?
@onvm can we try nn30?
Your results will arrive shortly
@onvm can we try nn30?
Error: ERROR: Failed to copy ONVM files to nimbnode30
@onvm why are permissions a problem?
@onvm why are permissions a problem?
Your results will arrive shortly
@onvm why are permissions a problem?
Error: ERROR: Failed to copy ONVM files to nimbnode30
Testing
Your results will arrive shortly
Testing
Error: ERROR: Failed to copy ONVM files to nimbnode30
@onvm with the new updates?
@onvm with the new updates?
Your results will arrive shortly
@onvm with the new updates?
Error: ERROR: Failed to fetch results from nimbnode17
@onvm that tiny error? thanks
@onvm that tiny error? thanks
Your results will arrive shortly
@onvm that tiny error? thanks
Error: ERROR: Failed to fetch results from nimbnode17
@onvm this shouldn't hang
@onvm this shouldn't hang
Your results will arrive shortly
@kevindweb Where the small things we discussed in the meeting updated? Ping me when its ready to merge (I think you're already working on mTCP so this is ready right?)
Testing
Your results will arrive shortly
@onvm Updated to latest version
This might or might not work @onvm
This might or might not work @onvm
Your results will arrive shortly
@koolzz I added information about benchmarks to the README and changed the name of the helper script symlink
@onvm Olá
@onvm Olá
Your results will arrive shortly
@onvm did the changes mess things up?
Finally have Pktgen running for our Continuous Integration in the nimbus cluster!
Summary:
Long awaited, and tested recently in another PR, the default run mode is now to run Pktgen for our base performance testing. Right now here's the basic steps, ci sends our worker node to reboot, then fires the worker script. This script, depending on the
worker-config
, runs a certain mode (right now just pktgen or speed_test). If pktgen mode (MODE="0"), then we run a script that calls the PktgenopenNetVM-Scripts/run-pktgen.sh
script in the other node through paramiko'sSSHCLient
. That uses our new lua script that sends packets for 30 seconds. We retrieve the data from basic monitor, send it back to CI for analysis. The new worker script also allows for multiple run modes, (MODE="0 1" for example). This way, we can potentially run pktgen then speed_tester and get all the results back to back after reboot.Usage:
Merging notes:
Please merge #143 because I merged that code into this one, and all those commits are cluttering up this commit log as you see below (27 commits). There should only be like 3 commits for this PR so far.
TODO before merging :
Test Plan:
We have to figure out how we want to do different modes, from Github comment parsing or something. This way, we can stress test that this works. We should also test that this works on nn30 with nn33 (pktgen) so we know it's scalable to new nodes we want to use.
Review:
@koolzz @dennisafa