IPAC 2MASS Working Group Meeting #109 Minutes

IPAC 2MASS Working Group Meeting #109 Minutes 11/26/96

Attendees: R. Cutri, T. Evans, J. Fowler, L. Fullmer, T. Jarrett, G. Kopan, J. Lampley, B. Light, S. Lo, J. Mazzarella, R. Scholey, S. Wheelock, J. White

AGENDA

  1. Network Reconfiguration

DISCUSSION

Network Reconfiguration

The IPAC network is being reconfigured with project-oriented machines grouped on sub-nets. This provides project-dedicated resources some isolation from public use and improves efficiency in data transfer between machines working on related tasks. This type of change has been planned for months, and now it has been implemented for the 2MASS machines karloff, kinski, lugosi, and rodan. S. Lo described the new configuration and outlined some of the questions that it raises.

As expected, this change has brought some problems to light, some new and some recurring old problems. For example, some adjustments have been made necessary by the fact that the 2MASS team members' home directories have been moved to lugosi, which is the server for the 2MASS subnet. It has also become necessary to eliminate access to /proj/2mass on the server blackhole, but a directory with the same name cannot be created on karloff or lugosi. A new directory on lugosi will be created; it will be named /2massp, and its subdirectories will include some of those currently under /proj/2mass. This solution is typical in that it is not difficult to do and involves only a temporary inconvenience. Solutions to most of the problems are similar, but in some cases, a significant part of the effort needed is just deciding which of several possible solutions to adopt. G. Kopan distributed a list of straw-man ground rules, requirements, and suggestions for some of the more important relocations.

One question is what to do with the "USER" directories under /proj/2mass. These were used for protocamera data analysis and have been used for some stages of 2MAPPS development. It is not clear that these directories are needed anymore, since /2massc/dev directories are available, along with partitions for storage of test data. But a lot of test environments would require extensive modification if the "USER" directories were eliminated, and apparently the utilization of these directories has taken very diverse forms, making a cure-all approach difficult.

The team leaned towards incorporating the "USER" directories in some format on lugosi, but considerations about system backups complicated the issue. For efficient backup procedures to take place, partition size limits must be enforced. J. Lampley distributed a list of the partitions on the four 2MASS machines that are currently backed up on a regular basis, along with the backup frequencies for each. Some progress toward appropriate partition assignment was made, but continuing work will be needed to achieve a final reconfiguration design.

It was decided later to increase considerably the space for home directories, as this space will be in the regular backup schedule, and has existing mechanisms for management. Special arrangements for additional backups may be arranged as needed.

Similar issues were raised regarding the /proj/2mass/bin directory. This contains many executables used by most of the team. Some effort will be needed to arrange for all the correct versions of these and other utility programs to be made available under the new configuration. It appears that these will be placed in /2massc/utl/bin on lugosi. The /proj/2mass/doc directory and its subdirectories will be put in /2massc/doc on lugosi.

Some protocamera databases will have to be moved or copied to tape. Some effort will be needed to ascertain whether anyone is still using these databases. T. Evans will research whether the previous main users of the largest of these databases still need access to them. To the extent possible, databases not known to be accessed anymore will be copied to tape and not maintained on disk.

It was decided that no changes will be made until after the December 5 deliveries and build of 2MAPPS 0.2.