MPC Status Page: Archive (2007 July-December)
This page describes enhancements to or problems that have occurred with the MPC and scripts and the fixes that have been made.Recent problems are listed elsewhere. Index of other older problems..
Older Enhancements and Resolved Problems
- Recently-Designated Vaisala Orbits in MPES
2007 Dec. 20: 10:30. A user reported that the MPES was not returning ephemerides for a number of objects designated since MPC preparation started. These objects have only Vaisala orbits and the rebuilding of some MPES orbit files during the MPC preparation process has caused these orbits to become unavailable to the MPES. The resumption of MPC processing mode in a few hours should make these objects available again. - DOU MPEC
2007 Dec. 17: 07:55. Last night's DOU MPEC failed due to two malformed orbits in one of the files. The offending orbits has been replaced and the bits of the DOU MPEC that failed are being rerun. - NEA*.DAT files
2007 Dec. 12: 10:50. The NEA*.DAT files available on the ftp site are truncated due to a corruption of a secondary data file used by the process that generates those files. The corrupted file has been replaced and the files should be corrected after the next DOU MPEC is issued. - DOU MPEC
2007 Dec. 6: 10:50. Last night's DOU MPEC failed due to a malformed orbit in one of the files. The offending orbit has been replaced and the bits of the DOU MPEC that failed are being rerun. - Access Problems?
2007 Dec. 6: 10:40. A number of e-mail reports have asked if we have problems with our cgi server. There are no problems internally, the webserver is responding to requests normally and quickly. Any access problems are caused by something between the MPC and the end-user. Our Computation Facility reports that there was a nine-hour outage yesterday at an observatory site about two miles from the main site, but they gave no indication that this outage affected the whole observatory. - Apparent Problem with AUTOACK TEL Lines
2007 Nov. 30: 21:00. A recent change in MPC preparation has required us to preprocess TEL lines when the batches are ACKed. A side-effect of this is that valid TEL lines are erroneously being flagged as incorrect. A new executable that removes this erroneous warning is now in place. - Possible Service Interruptions Nov. 28-30
2007 Nov. 28. Today we are starting some work on our cluster machines. We will bringing a new machine on-line (which will become the new webserver machine), configuring external shadow-set disk-storage devices and upgrading the old machines from VMS V8.2 to V8.3. This work will take place over the next several days. While the current webserver machine is being upgraded there will obviously be an interruption in cgi service. There may also be brief periods when cgi services are unavailable while other cluster machines are being upgraded.- Nov. 29: 10:30. The old cluster machine upgrades are starting.
- 15:10. The webserver upgrade is starting.
- 16:45. The upgrades are finished for the day. Automated processes, such as AUTOACK, are being restarted.
- Nov. 30: 08:35. The automated processes, including AUTOACK, have been paused, to allow the upgrades to continue.
- 18:25. The upgrades have been completed. Automated processes, including AUTOACK are being restarted. Configuration of the new webserver will begin over the weekend, but transfer of web services from the old webserver may take some weeks (a lot of data needs to be shifted and the move needs to be coordinated with the Computation Facility). Transfer of data directories to the new shadow sets will also begin this weekend, but this should have zero impact on internal and external services.
- Display Circular scripts reactivated
2007 Oct. 31: 17:00. The Display Circular scripts found on the pages of recent MPECs, IAUCs and CBETs have been reactivated. The Display Circular scripts on the pages of returned circulars are NOT reactivated, neither are the Display PS/PDF options on the IAUC circular pages. We are considering how best to update the static HTML pages with the outdated script calls and restore the missing functionality. - Webserver rebooted
2007 Oct. 17: 11:15. The MPC webserver rebooted itself automatically this morning. Normal service should have resumed. - Webserver sluggish
2007 Oct. 16: 13:45. The MPC webserver is sluggish, as a result of a large number of running jobs. We hope to have a new, faster webserver on-line within a month. - Network Problem?
2007 Oct. 13: 10:07. A problem has caused five computers of the cluster to reboot themselves just before 03:00. This is the same situation that occured on Oct. 5. AUTOACK was stalled until a few minutes ago. We believe normal service has been restored. - Network Problem?
2007 Oct. 5: 21:17. A problem has caused five computers of the cluster to reboot themselves. Since all are located in the same building (which is different to all the other cluster members) and all are on UPSes, it was presumably a network outage of some type. We believe that normal services were restored within minutes. - Two bug fixes
2007 Sept. 10: 08:30. Two long-standing bugs in services have been fixed. A problem affecting R.A. ordering of NEOCP objects at certain times of the year has been fixed, as has a problem with the non-appearance of certain comets (those with perturbed orbits published after the last MPC batch) not appearing in the MPChecker services. - Missing Observations in MPCOBS
2007 Sept. 3: 09:31. A user reported that observations of a specific object published in the most recent mid-month MPS batch were not available in the MPCOBS service. An examination of the report seems to show that all the observations from that batch are unavailable. We are investigating.- 10:02. The problem has been fixed. The mid-month MPS observations are available in MPCOBS.
- Missing Orbits in MPES
2007 Aug. 14: 11:26. A problem with the overwriting of one of the main pending orbit files with some old elements has caused some elements to be unavailable in the MPES. We are rebuilding the affected file.- 12:36. The file has been rebuilt. All the elements in this file will be republished on the next DOU MPEC.
- Upcoming Network Maintenance (August 9)
2007 Aug. 8: 11:23. We have been informed that Harvard will be performing maintenance on one of its gateway machines on Thursday, August 9. The work will begin at 05:30 EDT and should be completed by 07:00. During this maintenance window there will be a loss of connectivity into and out of the observatory. - Power Glitch at Observatory
2007 Aug. 7: 12:37. A glitch in the electricity supply at the Observatory appears to have fried one of our UPSs, with the result that three machines and some networking equipment lost power. We have moved equipment around, restarted the affected machines and everything seems to be back to normal. - Mid-month Observation Files in ECS
2007 Aug. 5: 10:21. Due to an apparent problem in shifting one set of cronjobs from a to-be-decommissioned CF linux box to a new linux box, the files of complete through the latest mid-month batch of MPS seem to have been built incorrectly. We have checked the logs from the VMS end and see no problem. We have no logs on the linux side where the problem occurred, so diagnosis may be difficult. We hope that the problem will be identified and fixed by the next mid-month MPS batch.- 10:47. Further checking shows that all the observations of numbered minor planets appear to be in the midmonth/mpn.arc file, but they are not sorted correctly. The midmonth/mpu.arc appears to be complete and sorted correctly.
- Checker files
2007 Aug. 1: 14:00. Due to a problem filing discovery information during the preparation of the July 30 MPCs, a number of secondary datafiles were built incorrectly, including some used by the MPES. We are rebuilding the affected datafiles.- 14:21. We believe all affected files have been rebuilt.
- Checker files
2007 July 26: 14:47. Following a Feedback comment, it has been determined that the procedure that updates the odd-epoch elements for use in the various checker services has been failing in recent days due to a diskspace problem. The services are functioning, but are simply missing recent orbits. We are attempting to restore the missing orbits.- 15:15. We think we have restored the missing orbits. The two cases that were mentioned in the Feedback comment seem to have been corrected.
- Comet magnitudes in MPES
2007 July 24: 20:50. A problem that caused comet total (m1) magnitudes to be output when fainter than the nuclear (m2) magnitudes has been fixed. - Minor Planet Designation script
2007 July 24: 11:54. A new version of the cgi script that is executed when one selects "Display all designations for this object" in the MPES has been installed (see the Mar. 30 status message). A version that will work on the Minor Planet Designations page will be forthcoming (it has to cope with a much wider range of possible inputs than the MPES version). - Upcoming Network Maintenance (July 26)
2007 July 23: 12:22. We have been informed that Harvard will be performing maintenance on one of its gateway machines on Thursday, July 26. The work will begin at 04:30 EDT and should be completed by 06:00. During this maintenance window there will be a loss of connectivity into and out of the observatory. - Mis-flagged objects for numbering
2007 July 1: 18:43. A number of objects are being erroneously indicated as being "flagged for numbering in the next full MPC batch", principally a couple of one-opposition NEAs! This problem may correct itself after tonight's DOU MPEC is issued.