Many people will be away tomorrow (Dan, Igor, Tom, Richard and myself at least) so this week's meeting is "virtual". I have put a few status updates below, please feel free to add any other items others should be away of, or post comments/questions.
Blocking ports 8000:8999 on glastlnx*.
Navid has set up offsite port blocking using IPTables on glastlnx05 to test if this works. Among other things this means that if you need to access http://glast-tomcat03:8080 from offsite you will need to use VPN or an ssh tunnel. The blocking seems to have had the side-effect of blocking access to these same ports from glastlnx05 itself, which makes the ~/tomcat/scripts/stop55 command hang. Onec this is fixed, and assuming no other problems are found, we will proceed to do the same think on all glastlnx* nodes next week.
Pipeline II status
Pipeline II is now able to submit batch jobs and receive the batch status updates by e-mail. We are very close to having a first "released" version which Bryson can try out. Dan is working on updating the "stream status" (which is basically what is shown on the home page of the pipeline web interface). He has decided to implement this as a Java stored procedure (apparently it is too complex to do reasonably in PL/SQL) but this involved first learning how to upload Java stored procedures into glast-oracle02.
Karen has succedded in encapsulating heirarchical queries into oracle "views" which will help making the output from the log viewer understandable.
New IIS servers for glast-ground
The windows group has got windows file replication working, so hopefully the new IIS servers are almost ready to go into production. Some documentation is available.
Beamtest Preparation
Karen is setting up beamtest database table(s) based on Ander's input, and will work on a web query interface next week. We need to clarify how the data will get into the database (from the pipeline?), and how we will find the files related to the beamtest runs (though the pipeline I database?).
Pruner/Peeler
There are several outstanding peeler/pruner issues, although little progress has been made this week since Tom and Igor are away. In particular:
- DATASERV-27@jira DC2 pruner front-end failed on job submitted by Hira (no TCUT, all runs, CPU time estimate came out negative!). I need to fix this, I think it was an inadvertent side effect of the change to the number of input from from the first to second version of DC2 data.
- Sylvain Guiriec's original problem with DC2 peeler was fixed (moved .so files). He subsequently asked if it would be possible to also support peeling MC files (deliberately left out of original DC2 peeler). Tom produced a new version of the peeler backend to support this, but an attempt to use it produced mysterious results (infilelist.txt generated by pre-release version of data server appeared to be unrelated to the requested task).
Misc
Rodney is said to be in possession of a "prescription" for solving the "slow windows boot when on wireless" problem. Ask him if you want to try it out.