From pmw@roe.ac.uk Fri Jun 11 21:16:06 2004 Date: Fri, 11 Jun 2004 16:03:29 +0100 From: Peredur Williams To: Nigel Hambly , Eckhard Sutorius , Harvey MacGillivray , Ian Bond , Mike Read , Bob Mann Cc: Clive Davenhall , Andrew Lawrence , Andy Adamson , Peter Shillan , John Taylor , Jim Emerson , Martin Hill , Mike Irwin , Peredur Williams , Malcolm Stewart Subject: WFAU WSA weekly meeting minutes, 11th June 2004 Minutes of WFCAM Science Archive meeting: 11th June 2004 ------------------------------------------------------------ ------------------------------------------------------------ Present: GPS, MCH, JDT, ETWS, IAB, MAR, PMW Apologies: NCH, AL, JPE, HMG, JMS, RGM DONM: 10 am, Friday 25th June 2004, plate library. Actions discharged: ------------------- ACTION: NCH & PMW to meet on Monday 7th June pm to review VDMT papers and other issues for the coming 3 weeks. DONE ACTION: NCH to send ETWS details of the PyMSSQL problems before leaving for cuba-libres y cervezas in the Balearics. DONE Actions partly discharged but continuing: ----------------------------------------- ACTION: ALL to review risk register and think of any new internal or external risks that should be documented in the risk register. Defered until top level plan is explicitly defined. PMW tabled the VISTA risk register as an example. ACTION: JPE to design and set up centralised VDFS web pages. Progressing. (there is now a link to it from our WSA Twiki). - continuing; thanks for progressing these. Actions carried forward from 04/06/04 meeting: ----------------------------------------------- None this week. Specific points and new actions: -------------------------------- Project management: PMW reported on VDMT meeting on June 8th: JMS and PMW participating from Edinburgh. Much discussion of consequences of IAB's departure and need at both WFAU and CASU for keeping documentation up to date. We also discussed the question of documentation for a PDR for V4: its value, impact on work in hand and timing. It was agreed to return to the question at the July VDMT meeting, when the WFCAM delivery and commissioning time-scales were would be known and VDFS user requirements agreed. PMW had drafted adverts and further particulars for recruitment of IAB's replacement and the new post agreed with the VEGA bid for circulation: the plan was to get the advert to AAS Jobs by June 15, published July 1, closing date July 31, shortlist August, appoint September, start as soon as possible thereafter. Networking: Nothing new this week. Hardware: PMW reported that Eclipse had replaced the external SCSI cabling for 'amenhotep' and that JNTD would be testing it with writing data to the disks. MAR reported that 'ahmose' had become unstable and that he got JDBC SQL errors like: "Time-out occurred while waiting for buffer latch type 2 for page (16:10598233), database ID 7.845". It has been re-booted. JNTD is checking the disks. Software: IAB reported: "Work is progress on the code to carry out the final step in generating data corresponding to merged sources, for ingest. A sticking point was the fact that input source data is encoded as binary files with different word sizes (ie data types) for the various corresponding attributes. In common coding practice, this is simply hard wired by declaring the variables to be of the appropriate corresponding type (eg short for "smallint", float for "real", etc). However, for our situation this needs to be determined at run-time. I have dealt with this by creating an abstract "DataChannel" class which channels a big array of bytes into the TableData classes that I have previously designed. A number of subclasses are then implemented which carry out the type conversions appropriate for a particular SQL data type. At run time, a factory function can then be invoked which instantiates a data channel object of the sub-calss appropriate to the SQL data type." ETWS reported: I'm looking through the programs Ian has written, to make myself familiar with them. Aside that, I've started to install mxODBC on ericht to run tests with some other (non Windows 2003) servers. MAR reported: Worked on the image access code (WCS/AST class files). Further investigated handling of long running queries, not too sucessfully, doesn't look possible to interrupt long running query using the JDBC/SQL Server combo. Fall back is to set a timeout which allows SQL server to kill the job. No meeting next week owing to RAVE consortium meeting; meet again 25th.