Reports 1-1 of 1 Clear search Modify search
DGS (General)
takahiro.yamamoto - 20:43 Friday 23 January 2026 (36212) Print this report
A stress test of the new camera server
As a stress test of the new camera server, AS, OMC_REFL, OMC_TRANS, and POS were moved from k1cam0 to k1cam2.
Now, 7 cameras (BS, PR2, PR3, and above 4 cameras) are running on k1cam2.
If some trouble happens on necessary cameras for your activities, please execute following commands and leave a log about rough date/time.
k1ctr*> ssh k1cam2
k1cam2> /kagra/camera/pylon-camera-service stop $(CAMERA_NAME)
k1cam2> /kagra/camera/pylon-camera-service disable $(CAMERA_NAME)
k1cam2> exit
k1ctr*> ssh k1cam0
k1cam0> /kagra/camera/camera-service enable $(CAMERA_NAME)
k1cam2> /kagra/camera/camera-service start $(CAMERA_NAME)
k1cam2> exit
 
If all cameras will be alive on Monday without any troubles, remaining cameras on k1cam0 will be moved to k1cam2.
Current CPU load of k1cam2 is ~3-4, so remaining 3 cameras on k1cam0 can be probably moved.
Comments to this report:
takahiro.yamamoto - 22:27 Monday 26 January 2026 (36217) Print this report
There was no trouble during last weekend, so I moved all cameras on k1cam0 to k1cam2.
Camera@OMC_TRANS was moved from k1cam2 to k1cam1 due to compatibility issue between a new pylon-camera-server and OMC_LSC guardian.
Now k1cam0 has no camera and we can start to upgrade it without any conflict.

-----
All cameras on k1cam0 was able to be moved to k1cam2 which is running as a new pylon-camera-server. The status of CPU status is good, with handling 9 cameras and ~6 as CPU load (8C/16T). From the view point of only CPU load, the remaining nine cameras on k1cam1could also likely be handled by k1cam2. Anyway, k1cam0 has no camera now. So we achieved a goal to start upgrading k1cam0 without any conflict. Though we need to work at the server room in the mine, upgrade work can co-exist with the commissioning works. The most serious barrier is probably weather in this week.

Camera@OMC_TRANS which had been moved to k1cam2 in klog#36212 was moved again to k1cam1 which is running as an old camera-server because there is compatibility issue and the implementation of OMC_LSC guardian. A procedure to take picture and a save path of taken pictures are different between the old camera-server and the new pylon-camera-server. So when camera@OMC_TRANS will be migrated to the new pylon-camera-server, OMC_LSC guardian must be also modified at the same time. Especially for the difference in the procedure to take picture, it probably requires "conuter" like implementation. This fact is highly incompatible with the current complicated implementation of the OMC_LSC guardian (TEM00 search).

Finally, I also implemented the status checker for the new pylon-camera-server and the tertiary camera server (updating the database of cam_ioc, preparing a new checker script for the pylon-camera-server, and preparing a new systemd service definition). Detailed implementation can be found in the installation manual of the GigE camera server JGW-T2516613. By the way, I noticed that the status checker script (/kagra/bin/CAM/camera_status.sh) for the old camera-server used on k1cam0 and k1cam1 didn't work properly because of the permission to read/write EPICS records and some trivial typos. This script is launched by admin account not controls, so EPICS records cannot be updated properly. I fixed a bugs on this script and now camera status is updated properly for all camera server.
Search Help
×

Warning

×