[Starlingx-discuss] the install progress information has missing when install computer node , storage node, and new controller node
Hi Expert; Issue descript : =================================================================================================================================== That greetings from intel NPG's Zhiwei, Now we try to using STX version of "stx-2018-11-13-nova-ibrs.iso", to enable computer node , ceph storage node etc Has a issue happens with "no any install progress information when the install node ", even install controller node, ceph storage node and computer node etc, fail , just show status is offline, It seem the install progress information has missing and progress status bar does not show on web page. Only have install success then will show "availability | online " The issue does not happen in titanium cloud 18.09. Issue logs: =================================================================================================================================== [root@controller-0 ~(keystone_admin)]# system host-show compute-1 +---------------------+--------------------------------------+ | Property | Value | +---------------------+--------------------------------------+ | action | none | | administrative | locked | | availability | offline | | bm_ip | None | | bm_type | None | | bm_username | None | | boot_device | sda | | capabilities | {} | | config_applied | None | | config_status | None | | config_target | None | | console | ttyS0,115200 | | created_at | 2018-12-27T13:54:56.393550+00:00 | | hostname | compute-1 | | id | 24 | | install_output | text | | install_state | None | | install_state_info | None | | invprovision | None | | location | {} | | mgmt_ip | 192.168.204.15 | | mgmt_mac | b4:96:91:0d:ee:0c | | operational | disabled | | personality | compute | | reserved | False | | rootfs_device | sda | | serialid | None | | software_load | 18.10 | | subfunctions | compute,lowlatency | | task | | | tboot | false | | ttys_dcd | None | | updated_at | 2018-12-27T13:56:36.260324+00:00 | | uptime | 0 | | uuid | ec03234c-29b6-4df4-b027-f4f51a4ffb85 | | vim_progress_status | None | +---------------------+--------------------------------------+ Thanks! zhiwei
+ Yan: is this something related to stx-gui? From: He, Zhiwei [mailto:zhiwei.he@intel.com] Sent: Thursday, December 27, 2018 3:34 PM To: starlingx-discuss <starlingx-discuss@lists.starlingx.io> Cc: Li, Baoqian <baoqian.li@intel.com> Subject: [Starlingx-discuss] the install progress information has missing when install computer node , storage node, and new controller node Hi Expert; Issue descript : =================================================================================================================================== That greetings from intel NPG's Zhiwei, Now we try to using STX version of "stx-2018-11-13-nova-ibrs.iso", to enable computer node , ceph storage node etc Has a issue happens with "no any install progress information when the install node ", even install controller node, ceph storage node and computer node etc, fail , just show status is offline, It seem the install progress information has missing and progress status bar does not show on web page. Only have install success then will show "availability | online " The issue does not happen in titanium cloud 18.09. Issue logs: =================================================================================================================================== [root@controller-0 ~(keystone_admin)]# system host-show compute-1 +---------------------+--------------------------------------+ | Property | Value | +---------------------+--------------------------------------+ | action | none | | administrative | locked | | availability | offline | | bm_ip | None | | bm_type | None | | bm_username | None | | boot_device | sda | | capabilities | {} | | config_applied | None | | config_status | None | | config_target | None | | console | ttyS0,115200 | | created_at | 2018-12-27T13:54:56.393550+00:00 | | hostname | compute-1 | | id | 24 | | install_output | text | | install_state | None | | install_state_info | None | | invprovision | None | | location | {} | | mgmt_ip | 192.168.204.15 | | mgmt_mac | b4:96:91:0d:ee:0c | | operational | disabled | | personality | compute | | reserved | False | | rootfs_device | sda | | serialid | None | | software_load | 18.10 | | subfunctions | compute,lowlatency | | task | | | tboot | false | | ttys_dcd | None | | updated_at | 2018-12-27T13:56:36.260324+00:00 | | uptime | 0 | | uuid | ec03234c-29b6-4df4-b027-f4f51a4ffb85 | | vim_progress_status | None | +---------------------+--------------------------------------+ Thanks! zhiwei
There are multiple components to the progress notifications. 1. The pxelinux.cfg files have a tisnotify=URL parameter that the patched Anaconda uses as a destination URL for notifications (look in /pxeboot/pxelinux.cfg/ files on the active controller) tisnotify=http://pxecontroller:6385/v1/ihosts/81adb329-0641-496e-acce-04a9d13cf865/ins... 2. Anaconda has a patch (stx-integ/base/anaconda/centos/patches/0001-TIS-Progress-and-error-handling.patch) that adds notification support. While your host is installing, you should be able to access the installer shell and check that the source code is patched, or check the content of the squashfs.img (/www/pages/feed/rel-19.01/LiveOS/squashfs.img) Example: controller-1:~/dpenney# ORIG_SQUASHFS=/www/pages/feed/rel-19.01/LiveOS/squashfs.img controller-1:~/dpenney# mkdir squashfs.mnt controller-1:~/dpenney# mount -o loop -t squashfs $ORIG_SQUASHFS squashfs.mnt controller-1:~/dpenney# mkdir LiveOS controller-1:~/dpenney# cp squashfs.mnt/LiveOS/rootfs.img LiveOS/ controller-1:~/dpenney# umount squashfs.mnt controller-1:~/dpenney# mkdir squashfs.work controller-1:~/dpenney# mount -o loop LiveOS/rootfs.img squashfs.work controller-1:~/dpenney# cd squashfs.work controller-1:~/dpenney/squashfs.work# ls usr/lib64/python2.7/site-packages/pyanaconda/tisnotify.py usr/lib64/python2.7/site-packages/pyanaconda/tisnotify.py controller-1:~/dpenney/squashfs.work# cd .. controller-1:~/dpenney# umount squashfs.work 3. Sysinv records the install_progress notification in the database as install_state, whether it's a string state or a number (percentage) 4. The dashboard reads and displays the install_state, with a progress bar if it's a percentage (stx-gui/starlingx-dashboard/starlingx-dashboard/starlingx_dashboard/dashboards/admin/inventory/tables.py) If your installer rootfs (squashfs.img) doesn't have the patch, you won't get notifications. Similarly, if the restructured dashboard isn't loading this properly, you won't see a progress bar. From: Xie, Cindy [mailto:cindy.xie@intel.com] Sent: Thursday, December 27, 2018 4:11 AM To: He, Zhiwei; starlingx-discuss; Chen, Yan Cc: Li, Baoqian Subject: Re: [Starlingx-discuss] the install progress information has missing when install computer node , storage node, and new controller node + Yan: is this something related to stx-gui? From: He, Zhiwei [mailto:zhiwei.he@intel.com] Sent: Thursday, December 27, 2018 3:34 PM To: starlingx-discuss <starlingx-discuss@lists.starlingx.io> Cc: Li, Baoqian <baoqian.li@intel.com> Subject: [Starlingx-discuss] the install progress information has missing when install computer node , storage node, and new controller node Hi Expert; Issue descript : =================================================================================================================================== That greetings from intel NPG's Zhiwei, Now we try to using STX version of "stx-2018-11-13-nova-ibrs.iso", to enable computer node , ceph storage node etc Has a issue happens with "no any install progress information when the install node ", even install controller node, ceph storage node and computer node etc, fail , just show status is offline, It seem the install progress information has missing and progress status bar does not show on web page. Only have install success then will show "availability | online " The issue does not happen in titanium cloud 18.09. Issue logs: =================================================================================================================================== [root@controller-0 ~(keystone_admin)]# system host-show compute-1 +---------------------+--------------------------------------+ | Property | Value | +---------------------+--------------------------------------+ | action | none | | administrative | locked | | availability | offline | | bm_ip | None | | bm_type | None | | bm_username | None | | boot_device | sda | | capabilities | {} | | config_applied | None | | config_status | None | | config_target | None | | console | ttyS0,115200 | | created_at | 2018-12-27T13:54:56.393550+00:00 | | hostname | compute-1 | | id | 24 | | install_output | text | | install_state | None | | install_state_info | None | | invprovision | None | | location | {} | | mgmt_ip | 192.168.204.15 | | mgmt_mac | b4:96:91:0d:ee:0c | | operational | disabled | | personality | compute | | reserved | False | | rootfs_device | sda | | serialid | None | | software_load | 18.10 | | subfunctions | compute,lowlatency | | task | | | tboot | false | | ttys_dcd | None | | updated_at | 2018-12-27T13:56:36.260324+00:00 | | uptime | 0 | | uuid | ec03234c-29b6-4df4-b027-f4f51a4ffb85 | | vim_progress_status | None | +---------------------+--------------------------------------+ Thanks! zhiwei
participants (3)
-
He, Zhiwei
-
Penney, Don
-
Xie, Cindy