IAU Minor Planet Center

MPC Status Page: Archive (2005 January-June)

Recent problems are listed elsewhere. Index of other older problems..

Older Enhancements and Resolved Problems

• Comments in the NEOCP
  2005 June 30: 08:10. A number of reports have been received in the past few days about the non-functioning of the Leave Comment form on the NEOCP. There was a brief problem with that form a few days ago, as a result of making the page HTML-4.01 compliant. Since few details were supplied in the reports, it was assumed that the problem lay in the use of locally-cached copies of the problem versions of the page. A report this morning suggested that this was not the case. Examination of the HTML code showed that while the template file for ToConfirm.html had been fixed, the template file for ToConfirmRA.html hadn't. It seems that the bug reports were referring to this latter page. The other template file has been fixed and both versions of the Leave Comment form should now be working. This exercise demonstrates the necessity of *detailed* bug reports if bugs are to be fixed promptly without wasting our time.

• MPCs in ADS
  2005 June 27: 22:00. After a gap of almost a year(!), the MPCs are again being indexed in the ADS. This problem that caused the ADS mail system to rejecting our e-mails containing author and title information has been fixed. The 2005 Apr. 7, May 23 and June 22 are included. When we figure out exactly which batches are missing, we will resubmit them.

• MPC Search facility
  2005 June 27: 14:00. A user reported (via an incorrect channel...) that the search facility was not working. This appears to be the case. We will take it up with the Computation Facility (CF) tomorrow, as they update the index for us.
  • June 28: 08:00. The problem was reported to the CF: they are investigating.

• New service: NEAObs
  2005 June 24: 12:30. A new service for observers of NEAs has been put on-line to allow external testing to occur. NEAObs is an observational planning aid that allows observers to selectively prune the list of currently-observable NEAs using a lengthy list of options. The service is currently to be considered beta code: bug reports to mpc at cfa.harvard.edu, please.
  • June 30: 23:40. As soon as the PDF documentation is ready the NEAObs service will cease being in beta test. However, the service is now up and running fully.

• HTML standards compliance
  2005 June 23: 23:15. Astute surfers will have noticed the appearance of HTML-4.01 compliance icons on a number of the MPC's pages. Although the majority of the MPC's pages have rendered well in the past on a wide range of graphic and textual browsers, we have been a little lax in strict adherence to the standards. Henceforth, as existing pages are modified or new pages brought on-line, greater effort will be made to ensure that the pages adhere rigorously to the W3C HTML 4.01 standards.

• Changes in MPES
  2005 June 23: 18:45. A slightly-modified PDF description of the MPES is now available. In order to better support a new service, the uncertainty information returned by the MPES is now defined by the semimajor axis of the uncertainty ellipse, rather than the major axis.

• Residual blocks in MPES
  2005 June 22: 16:15. While fixing a problem with an outdated residual block being returned by the MPES some 750 residual blocks for objects designated in the first four months of 2005 were accidentally blanked out. The residual blocks for these objects will be regenerated as and when their orbits are recomputed.

• Missing recent objects in MPES
  2005 June 22: 10:00. A user reported that elements for recently designated objects were not available in the MPES. We have looked into this and think we see the problem. If the elements are available tomorrow, then we'll know that our suspicions are correct and we'll be able to fix the problem for the future.
  • 17:40. It was realised that the problem could be fixed in a rather simple fashion and this has now been done.

• Network maintenance Saturday, June 11
  2005 June 6: 17.25. We have just been informed that the Computation Facility will be upgrading network components on Saturday, June 11, from 06:00 to 10:30 EDT. During this period there may be scattered outages in inbound and outbound traffic.

• Cluster problems
  2005 May 25: 14:30. A clusterwide problem of as-yet-undetermined nature has stalled most of the cluster machines. We are investigating. Until this problem is resolved, processing (and ACK'ing) of observation batches is suspended and scripts processsed on our web server may or may not function.
  • 16:20. The cluster has been restored to normal operation. It may take a little while for all applications to return to normal operation.
  • 16:45. We think all applications are now restored to normal operation. Tonight's DOU MPEC will be attempted, but the publication of the latest mini-batch of MPCs will be delayed until tomorrow.
  • 21:25. We spoke too soon. Apparently, the SMTP queue on SCULLY (AUTOACK) was not working properly. So no ACKs were sent and no batches made it into the processing queues. We have restarted the SMTP queues on this machine. The CF mailserver should automatically resend any failed deliveries shortly.

• Full user disk
  2005 May 15: 10:00. It seems that a userdisk on one of the cluster machines filled up completely overnight. There are routines that check for low disk space on specific devices, but for user disks these are currently invoked only when a user logs in: in a VMS cluster enviroment, this can happen very infrequently. In this particular case, the last interactive login was on Feb. 21. The checks are being added to a frequently accessed application in order to avoid this happening in the future. Some overnight jobs running on the affected machine will need to rerun.

• NEA.DAT files in MPCORB
  2005 May 4: 18:00. A user reported that the NEA.DAT and NEAcr.DAT files accessible in MPCORB are blank. The reasons for this are unclear: we will see if the problem repeats after tonight's DOU MPEC is issued. While checking this problem out, it was noticed that the files of NEA elements at epochs close to the present date were lacking many objects: this problem has been fixed and all the NEAs should be present following the next DOU MPEC.
  • May 5: 10:10. The files of NEA elements at epochs close to the present date are now complete. The NEA*.DAT files (and a couple of other derivative files) are still blank. Until this is fixed, users should download the NEA+00.DAT file (which is what observers should be downloading in any case).

• DOU MPEC
  2005 May 1: 12:20. Last night's DOU MPEC has been abandoned. Any files that get updated as a result of DOU issuance will not be updated. Normal schedule should be resumed tonight.

• Graphic images
  2005 Apr. 13: 21:1. Graphic images generated on the fly by the NEOCP and MPES were unavailable for a period today. The directory in which the generated images were stored has apparently not been cleaned out in some time. The old images are being deleted and a batch job will be set up so that in the future the directory is cleared out nightly.

• Computer clocks
  2005 Apr. 3: 08:15 EST. For some as-yet-unexplained reason, the clocks on our cluster members did not make the DST adjustment at 02:00 this morning. While we could manually alter the clock, we want to understand why the change did not happen automatically. Until this understanding is achieved, the times on circulars will be one hour fast on UTC.
  • Apr. 4: 16:20 EDT. After investigating, we know the reason why the clocks didn't change automatically, but we're unsure as to why this situation arose. We have changed the clocks manually while we dig further.

• Incorrect/Missing elements in MPES
  2005 Mar. 16: 08:15. A number of users have reported incorrect matching of elements to requested designations. Our internal routines appear to be fine: the MPES uses the same routines, so we're unsure at this point why there is a problem on the Web. It is possible that this is related to the recent shifting of logical disks onto a shadow set, the subsequent reboot of a cluster member and the possible erroneous assignment of logical to physical disk pointers. We are investigating.
  • 08:22. On trying the MPES again, we are now getting the correct elements. We are leaning towards a problem stemming from logical->physical disk name translation that was fixed automatically following the latest DOU MPEC.

• Bad Version Number error
  2005 Mar. 16: 08:15. A number of scripts started producing the above error overnight as a result of a logfile reaching its maximum allowable version number. The immediate problem has been fixed and a fix is being implemented to help prevent this occuring in the future.

• NEOCP
  2005 Mar. 3: 22:10. A number of problems (specifically, old object being posted on the NEOCP and errors being displayed during generation of new ephemerides) have been fixed. The first problem was due to one of the cluster machine having an old definition for a logical disk name. The second problem was due to the NEOCP ephemeris generation code containing a subroutine with the same name as a library routine.

• Unnumbered elements in the MPES
  2005 Feb. 24: 21:15. A problem was (apparently) fixed concerning the non-availability of unnumbered elements in the MPES during the latter stages of Minor Planet Circular preparation.

• The name of (6267)
  2005 Jan. 31: 10:00. While fixing some inconsistencies in the discovery file of numbered objects the name of (6267) was inadvertently changed. The name has now been restored to its correct form in our master copy of the discovery file and its on-line derivatives.

• Filesizes in the ECS MPCUPDATE service
  2005 Jan. 26: 23:00. The filesizes shown on various pages in the ECS following today's update with the 2004 Jan. 25 MPCs are too small and seem to bear no relation to the true sizes of the files. This problem seems to have have existed for at least two months.
  • Jan. 27: 00:35. A recent change in the ECS-generation procedure means that the method used until now to calculate the filesizes is no longer valid. The program that generates the various pages for the ECS has been fixed and the affected pages have been regenerated.

• DOU MPEC and mid-month MPS batch
  2005 Jan. 24: 00:20. It seems that the Computation Facility box that we use to transfer material to/from the MPC/CBAT/ICQ webpages/ftp space is not working properly. We can connect, but then the session just hangs. The problem apparently occurred around 18:00 and it was reported to the CF at 21:25 yesterday evening but we have, as yet, received no response. In light of these problems, tonight's DOU MPEC is cancelled and the issuance of the latest mid-month MPS batch is delayed.
  • 09:48. As of yet, we have had no response from the CF. In addition, it seems the Observatory is closed today due to the lingering effects of the recent blizzard. Who knows when this will get fixed?
  • 11:07. The problem has apparently been fixed.

• Numbered elements in MPES
  2005 Jan. 6: 12:25. It seems that there was a disk problem during the preparation of the orbit files for use in the MPES following last night's DOU MPEC. The consequence of this is that the elements for most numbered objects are currently unavailable in the MPES. We are hoping that this was a transient problem and are rerunning the affected command files.
  • 13:25. It appears that one of the disks on the webserver machine is failing. Weather conditions prevent us getting into the office to attend to the problem, so we are transferring the contents of the problem disk on to another disk on another cluster member.
  • 14:15. Access to the numbered elements has been restored. Residual blocks for the unnumbered objects are not yet copied over to the new disk.

• Elements for software packages
  2005 Jan. 3: 18:00. It seems that the 2004/07/14 epoch elements for numbered objects in the files of elements for software packages were not updated to the 2005/01/30 epoch. They have now been replaced in the internal file used to generate these files and should be become visible externally after the next DOU MPEC is prepared.