QA » openQA Project » openQA Tests

no capacity

Comparing with jobgroups in OSD:
i didnt found any that is used by others.

Comparing with TW the following are used by both which if i got it right we cant remove them from the test suite:
[ RAID switch_keyboard_textmode gpt yast2_gui mediacheck crypt_no_lvm remote_ssh_controller RAID0 select_disk installer_extended yast2_nfs_v4_server lvm_thin_provisioning autoyast_multi_btrfs lvm+RAID1 autoyast_gnome yast_no_self_update RAID1 btrfs+warnings RAID10 clone_system yast2_nfs_v3_client yast2_nfs_v4_client cryptlvm remote_vnc_controller nvme autoyast_reinstal autoyast_mini lvm-encrypt-separate-boot modify_existing_partition boot_linuxrc lvm yast2_nfs_v3_server xfs switch_keyboard_gnome yast2_ncurses ]

So the following can be used from the Tina's script for now:
'RAID5',
'RAID6',
'USBinstall',
'addon-module-ftp',
'addon-module-http',
'allmodules+allpatterns',
'autologin_yast',
'autoyast-supportserver',
'autoyast_bcache',
'autoyast_btrfs',
'autoyast_bug-872532_ix64ph1069',
'autoyast_bug-876411_btrfs_h5_autoinst',
'autoyast_bug-877438_ix64ph1029',
'autoyast_bug-879147_autoinst',
'autoyast_bug-887126_autoinst',
'autoyast_bug-887653_autoinst_jy-snapshot',
'autoyast_disk_as_md_member',
'autoyast_disk_as_pv',
'autoyast_error_dialog',
'autoyast_eula_full_medium',
'autoyast_ext4',
'autoyast_invalid_default_target',
'autoyast_keyboard_layout',
'autoyast_lvm',
'autoyast_multipath',
'autoyast_nfs_share',
'autoyast_non_existing_graphical_target',
'autoyast_reinstall',
'autoyast_resize_luks2',
'autoyast_tftp',
'autoyast_y2_firstboot',
'autoyast_zkvm_sles_product_reg',
'btrfs_libstorage-ng',
'cryptlvm+activate_existing',
'cryptlvm+activate_existing+force_recompute',
'cryptlvm+activate_existing+import_users',
'cryptlvm+cancel_existing',
'detect_yast2_failures',
'dud_development_tools',
'ext4_yast',
'gnome_http',
'gnome_self_signed_https',
'gnome_smb',
'installer_extended_textmode',
'installer_extended_x11',
'iscsi_client_normal_auth_backstore_fileio',
'iscsi_client_normal_auth_backstore_hdd',
'iscsi_client_normal_auth_backstore_lvm',
'iscsi_ibft',
'iscsi_server_normal_auth_backstore_fileio',
'iscsi_server_normal_auth_backstore_hdd',
'iscsi_server_normal_auth_backstore_lvm',
'lvm+cancel_existing_cryptlvm',
'lvm+resize_root',
'lvm-full-encrypt',
'minimal+base_yast',
'minimal+role_minimal',
'msdos',
'multipath',
'nis_client',
'nis_server',
'offline_install+skip_registration',
'releasenotes_origin+unregistered',
'remote_ssh_target_ftp',
'remote_vnc_target_nfs',
'repo_inst',
'skip_registration',
'sles+sdk+proxy_SCC_via_YaST',
'yast2_cmd',
'yast2_firstboot',
'yast2_ui_devel',
'yast_hostname',
'yast_hostname+dhcp_hostname',
'yast_hostname+linuxrc_hostname',
'yast_self_update',
'zfcp'}

And this list maybe it should be compared with other groups in o3 but
I think it would be nice to have a script or something to find them automatically in the future.

I have generated this list (see attachment) which i hope it can help to do the next step.

In the file, there are 208 combinations of pairs. I "invented" a matching rate for each one.

The calculations made by comparing and matching the key:values of each pair in an additional combination of the remaining matched keys(but not values)

Obviously, the findings are not accurate, in sense that this list provides some trivial investigation but it cannot be completed.

For example: the pair (remote_ssh_target_ftp , remote_vnc_target_nfs) have 0.82 match rate because they have [INSTALLONLY, NETBOOT, NICTYPE, OVS_DEBUG, USE_SUPPORT_SERVER, WORKER_CLASS]. But as i understand, this is a case that we cannot reduce.
In additional the findings can be reduced further because the number of the the combinations is with a pair of two. As an example: RAID pairs can be combined all in one.

I have attached another file which i tried to make groups of test suites, based in the previous list) that contains a set of common variables.
This turned to be harder than i thought so I spent sometime to collect some data to categorize the test_suites per the_most_used variables.
So for the first 10 most common used variables are:

[('1', 202),
 ('gnome', 48),
 ('tap', 38),
 ('textmode', 31),
 ('SLES-%VERSION%-%ARCH%-Build%BUILD%@%MACHINE%-gnome.qcow2', 15),
 ('0', 13),
 ('scsi-hd', 12),
 ('create_hdd_gnome', 11),
 ('14400', 10),
 ('SLES-%VERSION%-%ARCH%-Build%BUILD%@%MACHINE%-gnome-uefi-vars.qcow2', 8)]

where from them the No1, which shows the value that enables a variable, breaks down to (first 10 elements aagain):
[('INSTALLONLY', 59),
 ('BOOT_HDD_IMAGE', 22),
 ('AUTOYAST_PREPARE_PROFILE', 13),
 ('OVS_DEBUG', 10),
 ('LVM', 9),
 ('ENCRYPT', 8),
 ('EXIT_AFTER_START_INSTALL', 7),
 ('NETBOOT', 6),
 ('USE_SUPPORT_SERVER', 4),
 ('ENCRYPT_ACTIVATE_EXISTING', 3)]

I think from here we can think of a way to group tests more efficient. For example all the test with

Converting this to epic as we will have multiple subtasks for it.

due to changes in a related task

due to changes in a related task: #65393

due to changes in a related task: #66964

due to changes in a related task: #66964

due to changes in a related task: #66967

due to changes in a related task: #66967

Good job, team!