Status Report for SLAC/DOE IEPM Project
For January 2002
Work completed or in progress since the last report.
IPEXWarren worked with the administrator from Telcordia to install the IPEX monitor at SLAC. The installation went smoothly but security issues have delayed the start of data collecting.
IPMPWarren worked with security over their concerns for allowing the IPMP protocol to be tested with the SLAC AMP box.
NIMIWarren upgraded the NIMI monitor to a system that can be supported by the SLAC unix administrators.
SurveyorWarren powered down the Surveyor box. The project appears to be quite dead.
VoIPWarren worked with the central Stanford Network group to resolve a problem of packets being blocked to the SLAC VoIP gateway. Warren is also working with the telephone group to provide a recorded message for collaborators in Brazil to test the quality of their VoIP performance.
CertificatesWarren ran into problems configuring the SLAC web server to recognise the globus user certificate. Adil is installing the latest version of Globus so we can start testing GridFTP.
Web100Doug installed the web100 enabled iperf tool and is installing web100 on further machines. Doug is working with Jason Lee to get iperf running on Web100 and with Tom Dunnigan to port webd to SLAC. He is also writing a tool to faciliate extracting the web100 variables.
IPEP-BW / Grid-Monitoring
- New hosts and collaborators
- Les added new nodes for CERN, STARTap, IN2P3 (in order to look at perfomance via Renater). Les got an account for Guojun of LBL the author of pipechar. I want him to look at why pharlap often gives incorrect answers, and different from hercules.
- Ran into a reproducible problem with a Solaris host hanging when trying to copy a file from AFS to a file system mounted with no caching. Reported to Sun and responded to requests for further information, upgraded to latest release of Solaris, but still not fixed.
- Gary Buhrmaster, Connie and Les worked on poor monitoring performance following setting some trace debugging commands on pharlap see A Sudden Change in Bandwidth Utilization.
- Les had email exchanges with UWisc, UFL, NIKHEF, JLab UTDallas, LBL, GSFC, BNL and FNAL. to resolve port, ssh, disk space, or inaccessible problems
- Les, Connie Jerrod baby sat the monitoring to look for pathologies, come up with signatures, and Jerrod wrote a tool to assist in analyzing the logs to help identify problems. Les also formally started tracking problems. Several improvements were made by Connie and Les to the measurement program to improve reliablity, recognize more failiure modes, and better optimize the size of file to be copied
- Les spent a weekend in Washington at DARPA/DoE NGI joint PIs meeting. He gave presentation on PingER NG and wrote a Trip report. Les then spent another 3 days in Washington for the SciDAC meeting and gave another talk and wrote another trip report.
- Les responded to survey from Internet 2 End to End performance initiative on network measurements in preparation for the Tempe Arizona E2Epi meeting.
- Les also provided a quad slide and more information for INCITE and PingER for the SciDAC meeting in Washington.
- Les notified several mailing lists about the Passive and Active Measurements conference for which he is on the Program Committee. Working to get an account for Guojin of LBL the author of pipechar. I want him to look at why pharlap often gives incorrect answers, and different from hercules.
- Les & Warren attended 2 phone meetings with the PPDG monitoring group and 2 with the I2 E2Epi group.
- Les put together a presentation for the I2 HENP Working Group meeting] in Tempe Arizona.
- Les, Warren and Connie met with Brian Tierney, Guojun Jin and Greg Gurney of LBL to go over Net-100 and IEPM-BW status and plans.
- Les and Connie worked on defining a new simple prediction tool.
- Les documented the procedures for setting up new remote IEPM/BW hosts
- Les also documented the experiences with developing IEPM-BW.
Missing Data and New Sites
- The monitoring sites at BNL is still down and there is no estimate for configuring a replacement system. The site at Transpac has been brought back up. The web site serving the data from Imperial College was blocked by a firewall but is now open. Also there have been the usual glitches and minor loses. Les sent email to our Israeli HEP contact to try and get the ILAN PingER site working again.
- A new site at SISSA in Italy has joined the PingER HEP monitoring.
PeopleJiri Navratil from the University of Prague joined the IEPM group.
New SitesA new monitoring site at Trieste has been added.
PublicationsLes started making compression measurements for prospective publication at the Edinburgh GGF, as part of this he extended the automated measurement scripts, and drafted a web site on compression.
BaBar/HENP/PPDGLes provided assistance to a BaBar collaborator concerning using windows and streams. At the request of IN2P3 Les worked with ESnet and IN2P3 to try and understand the performance on the IN2P3-RENATER-ESNET-SLAC path, see Problems with performance between SLAC and IN2P3, Jan '02.
Les updated web page on network problem case studies for the PPDG monitoring group and the I2 E2Epi group.
Crated December 28, 2001.
Comments to firstname.lastname@example.org