Weekly Meeting

→ Europe/Zurich
CR04 (RAL)

CR04

RAL

Steve Fisher (RAL)
Description
<html> <head> </head> <body>
  • Past Events
    • SC06
  • Future Events
    • Copenhagen
  • Actions
    • Code Registry and Schema - Alastair - 27/4/2006
      Registry is just about there - but without replication of course. Uses a dummy schema. Using Martin's remote registry service which needed some changes. Not yet described in design.
    • Try to break JDK socket library - Alastair - 6/6/2006
      Various ways to get socket into closed wait state. 1.4 and 5 look similar. It would appear that JVM being out of memory can provoke this.
    • Follow-up on MonAMI - Antony - 5/7/2006
      Slight progress
    • Follow-up in Metadata Authz - Antony - 5/7/2006
    • Package monitoring scripts - Antony - 7/8/2006
      Generalized but not yet packaged.
    • Deploy monitoring scripts on PS,PPS and Preview - Antony - 7/8/2006
    • Investigate protection against running out of memory - Alastair - 7/8/2006
      He thinks that we should detect and restart the service - not clear how to do it. Lots of discussion... Maybe we should make some measurements. Alastair has done some work (it stops the service, trys to start the GC and restarts the service). Tested with big linked list on 1.5 - it did not recover well. Maybe add another exception to allow service to say it is too busy. Maybe do a better job of estimating total memory used.
    • Follow up on gEclipse - Steve - 2/10/2006
    • Consider extra functionality in SD to allow complex queries on ServiceData. - Paventhan - 2/10/2006
      Discussion has started. He has also started looking at the code and the SAGA documents.
    • Try ETICS (cleanly) - Alastair - 2/10/2006
      Done a bit and asked some questions. Alberto is converting an old 3.1 tag.
    • OGSA Naming into GLUE 2 - Antony - 2/10/2006
    • Service Proxy Module - add to GIN etc - Alastair - 2/10/2006
      Antony has done this for the job status raw archiver - can it be done in a more generic way? GIN has some bug fixes to make as well. Need to assign bug to Alastair - Antony will check that it is in Savannah.
    • Investigate Tomcat long running test (memory leak) and use of log4j - Alastair - 2/10/2006
    • Find why the PPS SFTs are not working - Antony - 16/10/2006
    • Watch the formation of the sensor group to make sure we are represented. Try to discourage key value pairs. - Steve - 16/10/2006
    • Write monitoring scripts to check state of XML pages for RTM - Antony - 13/11/2006
    • What has Stephen B taken over from Laurence? - Antony - 13/11/2006
    • Watch the R-GMA feed in L&B and submit a bug for FQAN etc. - Antony - 13/11/2006
    • Contribute to devlopers' guide. It should be small and HTML compatible - lots of links. No repetition of other documents. No attempts to cover up inadequacies of other documents. - Alastair - 13/11/2006
    • Contact Erwin concerning Copenhagen meeting attendance - Steve - 13/11/2006
    • Add FIXMEs for the schema design - Steve - 13/11/2006
    • Circulate list of mail lists for people doing support - Alastair - 13/11/2006
    • Update web page with instructions for how to carry out support - Antony - 13/11/2006
    • Create new support rota - Steve - 13/11/2006
    • Track down what has gone wrong with GridView reports. - Alastair - 13/11/2006
    • Consider jGAF - ??? - 17/11/2006
    • Study - mode/state in consumer maybe mode is a substate. We should try to eliminate modes that overlap with state information- ??? - 17/11/2006
    • Study start() in Consumer. In what circumstances can mode not be ``STARTED'' but already be RUNNING. Worse still if the task starts too soon (before STARTED) it will just do nothing. - ??? - 17/11/2006
    • Eliminate RemoteService and RemoteResourceService - Kush - 17/12/2006
    • In the Consumer the maximum count of tuples to pop does not appear to have been coded. This needs checking.- ??? - 17/11/2006
    • The consumer has code to remove things from the HashMap (m_replies) but does not appear to add them. This needs checking.- ??? - 17/11/2006
    • removeProducer call in consumer. This call is not in the spec nor in the design! Currently for a directed query it just sends reply.abort() and unlinks the reply otherwise it makes an asynchronous call to the mediator to get update instructions which are then sent to modifyPlan which will in turn abort the reply and unlink them. This needs checking.- ??? - 17/11/2006
    • In the consumer wouldn't it be better to pass the TupleQueue reference to the ProducerReply - but would there be synch issues?. This needs checking.- ??? - 17/11/2006
    • Does the resource need a flag to show when ready? This flag may not be necessary - it needs some thought. If startup and shutdown are synchronous it would appear not to be needed. This needs checking.- ??? - 17/11/2006
    • In the resource code the resource manager is a singleton so there is only one timer therefore calling cancel will cancel all timer tasks. This needs checking.- ??? - 17/11/2006
  • LCG - Antony
  • PPS - Antony
  • Bugs
     #8099 Archivers are inconsistent awilson burke
     #14551 cont+old query times out with no result fisher sburke
     #16140 RGMA queries are slow fisher sburke
     #17312 flexi archiver needs sysV init script. aduncan lfield
     #18006 Continuous+old needs fixing to support backwrads compatibility in the future aduncan fisher
     #18487 R-GMA should be able to use default proxy locations awilson fisher
     #19500 Inspector Gadget does not allow entry of a port number for querying MON boxes mjiang aduncan
     #19645 The Sub exception org.edg.info.ReadTimedOutException is an inappropriate type fisher aduncan
     #19883 The resilient consumer died due to "rgma.RGMAException: Unable to locate an available Registry Service" awilson awilson
     #20090 tomcat spawning processes without any limit aduncan harenber
     #20358 2 INSERTs with >1min in between fail aduncan harenber
     #20788 JobMonitor table is empty awilson fisher
     #20977 Inspector Consumer Status Details page does not show memory usage mjiang aduncan
     #21113 rgma overwrites configuration script without saving old one aduncan diana
     #21382 client hostname is incorrect in ServiceStatusDetails mjiang awilson
     #21558 tomcat on MON box wearing down aduncan maart
     #21704 Under load of queries, RGMA server becomes irresponsive and dies aduncan delgadop
  • Support rota
    • 20/11/2006 - Alastair
    • 27/11/2006 - Ming
    • 4/12/2006 - Paventhan
    • 11/12/2006 - Kush
    • 18/12/2006 - Antony
  • Design FIXMEs - see how far we get
    • Remote Call Subsystem
    • Request Management Subsystem
    • Introduction
    • System API
    • Query answering subsystem
    • Parsing and validation subsystem
    • Operation Authz
    • Logging
    • Physical Installation
  • Plan for week ahead
  • AOB
</body> </html>
list of actions
Minutes
The agenda of this meeting is empty