Olivier Callot 5 June 2007 Electronic Logbooks n
Olivier Callot 5 June 2007 Electronic Logbooks n Why moving away from Atlog ? n Proposed improved configuration n Release schedule
Why moving away from Atlog ? u. Elog is the original software n Developed and maintained at PSI by Stefan Ritt n Tested in summer 2005 by Tobias Lohe, summer student. l Limitation u. Atlog n Use if over 100, 000 entries… is an Atlas modification of Elog Oracle as storage for entries, instead of the file system l Remove the performance limitation for big logbooks n It has been installed at the Pit with two servers, one for inside the LHCb pit network, one for outside access l Many logbooks pre-installed l Some already used, with few (<<100) entries Olivier Callot Electronic Logbooks 5 June 2007 2
u. But Atlog has several limitations n Can not import/export logbooks → impossible to merge Velo old logbooks used in ACDC and in construction n Can not delete or modify an item n Crash when an unknown command is sent l Doesn’t react properly to port scans, … n Attachments still as files, not in Oracle. n Doesn’t support accented characters n And limited response from the maintainer… u. Try to go back to plain Elog n Learn how to configure it for the current use l Commissioning logbooks don’t have the same use as official logbook of a running experiment, in particular in the number of entries per day, the automatic insertion of entries and the automatic insertion of run and fill numbers Olivier Callot Electronic Logbooks 5 June 2007 3
Proposed improved configuration u. All current (pit) and external (Velo, NIKHEF) logbooks will be available together n. I managed to have an ad-hoc modified Atlog service to give me back the news in Oracle, and understood how to get the Velo logbooks from both here and NIKHEF. u. I would like a simplified layout for the logbooks n Simply date, author, “system affected”, subject and text n All pit logbooks with a similar structure l Makes u. Use the maintenance easy user registration n Needed only to post an item, not for reading n Self registration: l Need a short name, typically your local account, your full name to appear in the Author field, and a dedicated password Olivier Callot Electronic Logbooks 5 June 2007 4
u. Can we avoid user registration ? n Yes, but then the author is free field, can contain unclear values l Who was “John” on 12 May 2005 ? n Entering your name + password once is not very different from entering it (alone) with each posting l Remember: n The authentication is not needed for reading browser remembers your name for 8 hours (can be changed) l Idea: The shift console will run from a generic account, but the shift crew entries should have a meaningful author field… u. A good way to have some command restricted n The concept of ‘administrators’ is implemented in Elog, they can add/remove users, reset passwords, edit configurations, … u. We could also protect write actions from outside the pit by password… But is it worth? Olivier Callot Electronic Logbooks 5 June 2007 5
Proposed schedule u. You can play with the proposed layout n http: //lbweb 01. cern. ch: 8081 l Not n Check is accessible only from inside CERN from outside, due to the CERN firewall settings… for unused logbooks and configuration l RICH 1 and RICH 2 or a single RICH? u. Need to disable all logbooks (Pit + Velo + NIKHEF) for conversion n The operation is still a bit tricky, adding new entries at the same time would not make it easier n Then restart the new service at the pit, and discontinue the Velo + NIKEF ones. u. Either Olivier Callot early next week, or from June 25 Electronic Logbooks 5 June 2007 6
Future u. Keep Write, Edit, Delete access inside and outside n Except if you disagree for your logbook. Let me know n This seems the most useful during the commissioning period u. At some point, allows only read access from outside n Write only possible from inside the LHCb Pit network n What we want during running period, no interference with the shift logbook! u. Settings and configuration of the Shift logbook n Want probably Fill and Run number in the entries n Make sure the author name is correct for shift crew entries n Automatic entries from LHC states, Run control, alarms, … Olivier Callot Electronic Logbooks 5 June 2007 7
Feed back needed u. This u. We is a very important tool for an experiment have (still) time to play with various settings n And try to implement improvements and new functionalities u. A shift logbook will probably look different from the commissioning one n Several entries per hour, not one per (few) days as now n Many short entries, now long reports of progress n But the same tool can probably do it. u. Thanks Olivier Callot for feed-back… Electronic Logbooks 5 June 2007 8
- Slides: 8