Adding intel source vars line to netcdf install script. Also changing hard coded version to latest, which is a symbolic link so should always be useable. larson-group/sys_admin#738 (details)
Making check to see if baseline output exists before trying to delete. (details)
Adding a sed command to the budget_balance bash script that will always set the l_lmm_stepping flag to false, since that flag does not currently give balanced budgets. (details)
Changing iptables script to look at the correct files for the new computer configuration. (details)
Updating iptables_patrol.bash script for Pop_OS! log file locations. (details)
Moving old_carson's apache comfiguration files to subfolder incase we need to access them again. (details)
Modifying apache_backup.bash to point to the new carson apache2 configuration files for backup. (details)
Updating rc.netfilter to work with new carson. (details)
Changing deploy script sys_admin configuration to have the correct iptables save sommand for new carson. (details)
Adding comma at end of config json to allow it to work. (details)
Final change to the deploy script sys_admin config fixing the post steps for all computers to allow full through run. (details)
Making it so we only use commits tracked back from a single parent tree. (details)
Changing git checkout command to git reset --hard, since that will wipe away changes caused by run commands if any are present. larson-group/sys_admin#763 (details)
Add a few words to make it easier to find unresolved commits. (details)
Adding stephens_2022_paper branch to script. (details)
Updating NetCDF paths to match the current version on Carson. #763 (details)
Adding functionality to recognize empty output directories, and also ignore failed runs when updating the baseline commit. larson-group/e3sm#38 (details)
Changes "components/cam" to "components/eam" (details)
Making it so that when the test exits because the baseline and final commit are the same, it does not result in an error when the unresolved list is empty. Before, it resulted in an error no matter what. (details)
Updating iptables_patrol.bash to actually count the Dropping messages. (details)
Correcting the file we are checking for Pop OS (details)
Updating driver script for testing on python 3 version of checkBudget.py (details)
Moving -j8 option to correct make call. Also changing mpi path from ending with _intel to ending with _ifort, since this is the correct name specified when installing mpi using the intel compiler. (details)
Added backup_all_scripts_driver.bash script that has been being run on carson to the sys_admin repo #694 (details)
Updating larson-group.sh script to support newer pgi version and the intel mpi version #786 (details)
updated larson-grtoup paths file so oneapi mpi compiler works properly #786 (details)
updated the dynamic_ip script to python 3 (details)
updated the dynamic_ip script to python 3 (details)
updated broken url finder script to python 3 #693 (details)
updated test_bacula_restore.py script to python 3 #693 (details)
changed iptables patrol script to no longer not be able to find dropping messages #787 (details)
In python range(1,1) will cause the loop not to run at all, so when we only had 1 commit to check (1 commit on the previous day) the baseline would never be updated. This is just a bug, so we want to start the range at 0 and subtract 1 everywhere we see i to compensate, ensuring that the loop runs but that the elements checked remain the same. (details)
Updated dynamic_ip.py script to no longer hold onto users when the ip changes. I tested it if fairly rigorously, and it always deletes hosts when they are removed from hosts file, adds when they are added to the hosts file, and properly updates iptables rules when the host ip changes. #798 (details)
I just realized how outdated the hosts file is on sys_admin, so I am adding the updated one that is currently on malan. (details)
updating host repo update scripts to support commit logging from clubb #797 (details)
There was an branch that has been updated by this script for awhile, but has not been commited to the sys_admin repo. I am commiting this change now. (details)
Making fixing change to make autoupdate message maker work properly (details)
Added comments to git updates scripts effected by autocommit message maker changes (details)
Added all python test suite configuration files to sys_admin to be (details)
Added short bash script utility to add profile.d path updates to .bashrc (details)
these are edited instal files for mpich and netcdf overseen and changes reccomened by gunther (details)
Adds Gunther's openmp branch to clubb_release repo (details)
Removing files we don't need, and making changes in carson's apache config to facilitate hosting both the larson-group site and jenkins. See https://github.com/larson-group/sys_admin/issues/828\#issuecomment-2125606960 for more detail. (details)
Adding intel source vars line to netcdf install script. Also changing hard coded version to latest, which is a symbolic link so should always be useable. larson-group/sys_admin#738
Adding a sed command to the budget_balance bash script that will always set the l_lmm_stepping flag to false, since that flag does not currently give balanced budgets.
See https://github.com/larson-group/clubb/issues/893#issuecomment-905712817.
Changing git checkout command to git reset --hard, since that will wipe away changes caused by run commands if any are present. larson-group/sys_admin#763
New carson has version 4.5.3 or NETCDF Fortran installed which may be preventing the CAM nightly tests from properly running. I've updated it from 4.4.5 to the proper version for new carson. #763
Making it so that when the test exits because the baseline and final commit are the same, it does not result in an error when the unresolved list is empty. Before, it resulted in an error no matter what.
Updating iptables_patrol.bash to actually count the Dropping messages.
Previously for the computers with a week long iptables check they were either not working correctly or grepping only one day. I added in a for loop to count through each day and add the result of each day together.
Moving -j8 option to correct make call. Also changing mpi path from ending with _intel to ending with _ifort, since this is the correct name specified when installing mpi using the intel compiler.
In python range(1,1) will cause the loop not to run at all, so when we only had 1 commit to check (1 commit on the previous day) the baseline would never be updated. This is just a bug, so we want to start the range at 0 and subtract 1 everywhere we see i to compensate, ensuring that the loop runs but that the elements checked remain the same.
Updated dynamic_ip.py script to no longer hold onto users when the ip changes. I tested it if fairly rigorously, and it always deletes hosts when they are removed from hosts file, adds when they are added to the hosts file, and properly updates iptables rules when the host ip changes. #798
Making fixing change to make autoupdate message maker work properly
initially i had written these with python as the system to run the code. On malan, this was python version 3.9, on carson, it was python 2. This change will make it work properly
Removing files we don't need, and making changes in carson's apache config to facilitate hosting both the larson-group site and jenkins. See https://github.com/larson-group/sys_admin/issues/828\#issuecomment-2125606960 for more detail.