• Processing (Info)

    • MPC Status Page: Archive (2004 January-June)

    This page describes enhancements to or problems that have occurred with the MPC's webpages and scripts and the fixes that have been made.

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

    Older Enhancements and Resolved Problems

    • Minor Planet Checker
      2004 June 29: 21:30. A user reported that the Minor Planet Checker was not finding some objects. Some objects were apparently missing from the check files, as a result of a problem with the generation of those files by the DOU MPEC. The generation routine has been recoded to try and eliminate the file-locking problem that caused the current problem. In the meantime, the files have been regenerated.

    • MPES
      2004 June 25: 10:40. A user reported that requesting a 90-date 50-day-interval ephemeris returned the ephemeris header, but no ephemeris data. There is a documented limit of 4000 days for the maximum extent of any one ephemeris. When requests exceed this limit, the request is to be truncated to the number of dates that will fit in 4000 days. This truncation was not being done correctly. The code has been fixed.

    • Minor Planet Checker
      2004 June 24: 17:00. A user reported that the Minor Planet Checker was apparently using out-of-date orbits. An check showed this to be the case. The MPChecker uses 40-day epoch elements to improve the accuracy of the displayed positions. A program that merges the 40-day elements from new orbit determinations with the 40-day elements from earlier determinations created a temporary file that overflowed the maximum allowable file version number. The command procedure has been fixed to prevent this happening in the future. All perturbed orbits published since June 12 will be republished on the next DOU MPEC to allow the new elements to become available in MPChecker. It is probable that the June 21 problem with MPChecker was related to this current problem.

    • Duplicate MPEC
      2004 June 24: 10:00. Due to problem with the CF boxes (on which the diskspace for the MPC's webpages resides) yesterday evening, MPEC 2004-M44 was issued twice. The DOU MPEC will remain MPEC M44, the new object will be reissued.

    • Minor Planet Checker
      2004 June 21: 11:00. A user reported that the Minor Planet Checker was apparently not returning matches with comets. A quick check showed this to be the case. The program that rebuilds the file used by the Checker was rerun (it is normally run once a day after the DOU MPEC is prepared). The Checker then returned matches to comets. The reason for this difference in behavior is not clear at the moment.

    • Network disruption
      2004 June 10: 13:00. We have been informed that CF will be installing security patches on the network switches on Friday June 11 between 08:00 and 11:00 EDT. This will certainly affect Internet connectivity and will probably cause our cluster to fail. We will try to resume normal service as soon as possible after the security patches are installed.
      • June 11: 10:30. The security patches have been installed and the cluster reformed. Normal service should have been resumed.

    • Missing designation messages
      2004 June 9: 22:00. We have recently implemented significant further automation in the processing of incoming observation batches. Observations of numbered and multi-opposition objects in incoming batches are processed, checked and new orbits redone as necessary, and orbits and observations are filed without any human intervention. Similarly, automated routines now perform the first-level check on batches of unidentified objects, extract identifiable objects and submit to the aforementioned automated routines and send the file of designations back to the observer. Unfortunately, we have just realised that this processing is being undertaken on machines where outgoing SMTP e-mail is blocked by the Computation Facility (see earlier discussions on this). Until we get this sorted out with the CF, observers will not receive files listing linkages of their objects with numbered and multi-opposition objects. We will attend to resending designation files from unblocked machines as time permits.

    • Network outage
      2004 June 7: 10:00. There was apparently a major network outage on the CfA network. Most of the clustered machines rebooted as they were unable to "see" other cluster members.
      • 12:00. All the affected machines have been rebooted and the cluster has been reformed.

    • MPCORB mirror site
      2004 June 3: 13:00. A user reported (to a staff member rather than to the generic mpc@cfa address...) that the Czech Astronomical Society ftp server mirror of the MPCORB datafiles had not been updated since May 26. It seems that the hardware for the ftp server was upgraded on this date, which had the net result that the SSH public key we had for this machine was no longer valid. We have a new public key, but there are still problems with running the update procedures. We are working with the ftp server's system manager to correct the problems.
      • June 6: 11:30. Thanks to Martin Kolarik, system manager at the remote ftp server, the server is again accessible to our update routines. The current data is being copied now, normal updates should resume with the next DOU MPEC.
    • Incorrect mean anomaly in some element files
      2004 May 30: 00:01. A user reported that the values of the mean anomaly for orbits in certain elements files prepared for various software packages all had the same value. This problem was caused by the failure to change the name of one library function call in code newly ported from the VAX to the Alpha. This has been fixed.

    • Planned network outage
      2004 May 25: 17:40. We have been informed that Harvard will be doing some maintenance work on their network which may affect our connection to the outside world. The one hour estimated maintenance window will start on June 1 at 06:45 EDT.

    • MPES
      2004 May 15: 12:40. A user reported that the MPES is returning orbits for the wrong objects. This appears to be a problem with a change in the way we maintain orbit elements internally and the problem is affecting us internally as well.
      • 12:50. The problem was identified and resolved.

    • Planned network outage
      2004 May 10: 10:14. We have been informed that Harvard will be doing some work on the router that connects the CfA to the outside world. The one hour estimated maintenance window will start on May 13 at 05:00 EDT.

    • DISCSTATUS reports
      2004 May 6: 16:20. The latest DISCSTATUS reports incorrectly report the named-status of the newly-numbered minor planets above (80000). No names have yet been assigned in this number range. A correction has been made to the DISCSTATUS program to prevent this happening in future months.
      • May 7: 11:00. This month's DISCSTATUS reports are being regenerated as yesterday's delays/problems caused some errors in the reports.

    • MPC preparation delays and "missing" objects in the MPES
      2004 May 6: 11:15. There have been a number of delays in the preparation of the latest MPC batch. One visible consequence of these delays is that elements of newly-numbered minor planets are not available in the MPES under either the unnumbered or numbered designation. This problem will self-correct during the day as we complete MPC preparation.
      • 14:51. The elements of the newly-numbered minor planets are available under their numbered or unnumbered designations.

    • Missing objects in MPES
      2004 May 5: 16:25. A user reported that a number of NEOs are missing from the MPES. An examination of the log file from last night's DOU MPEC shows that there was a problem with the generation of the files used by the MPES. We cannot regenerate these files at the present time as we are still preparing the latest batch of MPCs. The problem should be corrected after the next DOU MPEC is prepared.
      • 17:00. It proved possible to update the files needed by the MPES.

    • Dates of Last Observation pages
      2004 May 5: 10:35. A user reported that, although these pages are being updated, no details of new observations are appearing. This is a problem related to the MPC being in "Minor Planet Circular preparation mode". A fix has been put in place to ensure continued updating during future MPC preparations.

    • Ephemerides on SpaceJunk page
      2004 May 5: 10:34. Ephemerides for a number of objects on the SpaceJunk page are not available for current dates. This is a result of the ephemeris data for these dates being unavailable from the Satellite Situation Center (SSC). Ephemerides will be available again when the SSC updates their predictions.

    • IMP8 ephemeris on SpaceJunk page
      2004 Apr. 22: 10:22. A user reported that the SpaceJunk page was not returning an ephemeris for IMP8. This has been fixed.

    • Daily element files
      2004 Apr. 20: 11:10. A user reported that the DAILY.ELS file in the ECS was not updated. A quick check showed that the file was generated correctly and copied to the webserver, but a cronjob that was supposed to copy the file to its final destination and copy it to the mirror sites was not run. It seems that this cronjob has not been run since one of the CF's boxes was rebooted on Apr. 18. We are trying to track down the problem with the CF. The file in the ECS is now correct.
      • We also have reports of other files (e.g., neotod.dat and the mirror sites) that have not been updated. Again, this is a problem with cronjobs on the CF boxes.
      • 11:35. We have now heard back from the CF. Apparently, after the reboot on Sunday, the cron jobs did not get restored correctly. Cron was restarted around 11:00.

    • NEA*.DAT files on anon-ftp server
      2004 Apr. 10: 13:00. A user reported that the NEA*.DAT files on the anon-ftp server were missing the bulk of objects with designations after 2004 DA53. A change has been made to the routine that generates these elements and the files are being regenerated.
      • 13:30. Updated files are on-line.
      • 15:55. It turns out the updated files were put into a temporary directory and the script to copy them to the proper destination was not run. This is now being run. Note that the mirror sites will not get updated until the next DOU MPEC is run.
      • 16:30. Updated files are on-line in the correct directory.

    • MPChecker files
      2004 Apr. 7: 11:00. A user reported that MPChecker was not identifying objects. A check showed that the files used by MPChecker had been truncated. The affected files have been replaced. A change has been made in the routine that copies the files for use by MPChecker to prevent zero-length files being copied to the web server.

    • Sky Coverage plots
      2004 Apr. 6: 11:18. A user reported that requesting fields over a range of dates would cause the fields for the earliest date to be omitted. This has been fixed.
      • 11:44. While checking out the modified Sky Coverage script, it noticed that requesting plots showing (or excluding) fields shot by a list of sites only included (or excluded) the last site specified. This has been fixed.
      • 11:55. While checking out the previous fix, it was noticed that the "just big surveys" and "just others" code had never been activated. This has been fixed.

    • MPChecker
      2004 Mar. 24: 10:57. For objects with perturbed orbital elements, MPChecker now uses elements at the nearest 40-day epoch date rather than at the standard 200-day epoch date. This will help improve the accuracy of the "quick look" positions, especially for NEOs.

    • Mid-month-MPS-complete batches of observations in ECS
      2004 Mar. 23: 17:00. A user reported anolmalously large sizes for for the complete-through-the-latest-MPS-batch files of minor planet observations. This problem is related to the March 21 diskspace problem and was exacerbated by the non-logical functioning of a Unix utility. The files are being regenerated.
      • 18:52. The unnumbered file has been regenerated.
      • 21:47. The numbered file has been regenerated.

    • Planned radar observations at Goldstone
      2004 Mar. 22: 14:50. The MPES now flags those objects that are scheduled (or have been requested) for observation by radar from Goldstone. Comments are made as to whether astrometry or physical studies of the objects have been requested.

    • Display MPEC script
      2004 Mar. 22: 12:15. A user reported a problem with the form at the bottom of the Recent MPECs page, specifically that one could not access circulars numbered above 99. The script has been updated to fix this problem.

    • MPEC orbits in MPES
      2004 Mar. 22: 11:00. A user reported a problem with the orbit of 2004 FY3 downloaded from the MPES shortly after the discovery MPEC was issued. Instead of the perturbed orbit used on the MPEC, an unperturbed orbit was returned. This turns out to be a problem with the MPEC production code and it would only appear if a particular sequence of orbit calculations occurred for the object being MPECed. The problem code has been located and tightened up to prevent this happening in the future.

    • March 21 DOU MPEC and Mid-Month MPS batch
      2004 Mar. 21: 10:10. One of the cluster-shared disks filled up late last night. Remedial action (purging the disk) was not undertaken at that time. As a result the DOU MPEC failed to run. It has been abandoned.

      The problem also affected the preparation of some files associated with the mid-month MPS batch. The affected files are being regenerated.

      • Regenerated files are now available.

    • Erroneous line on today's MPEC
      2004 Mar. 19: 11:57. The four MPECs issued today have an erroneous line after the observer details on the copies that were mailed out. The MPEC production has been modified to prevent these lines appearing in the future.

    • 2004 FH orbit on DOU MPEC
      2004 Mar. 18: 10:40. The orbit for 2004 FH published on the latest DOU MPEC had an internal flag set incorrectly. This caused the problems reported by some users. A new orbit appears on MPEC 2004-F26 that corrects the problem.
      • 16:06. A user reported a problem with generating a 1h-interval ephemeris starting on March 16. This appears to be a problem with an internal pointer getting confused by the use of a non-standard epoch for the 2004 FH orbit. This may take time to fix as it will require examination and modification of a library routine. Use of the MPES for current ephemerides is not affected by this problem.
        • 17:12. A fixed version of the code, linked to the modified offending library routine (one of those WTFIsThatDoingThere? moments, in a bit of code called perhaps once in a million runs!), is now on-line.

    • Special-Epoch NEA element files
      2004 Mar. 17: 10:40. A user reported that the special-epoch NEA element files on the ftp site contained only one set of elements. The reasons for this are unclear. The files will have to be regenerated from internal copies.
      • 12:20. After making some changes to internal procedures, regenerated versions are now available.

    • MPChecker files
      2004 Mar. 15: 10:00. A user reported that MPChecker was not identifying objects. A check showed that the files used by MPChecker had been truncated. The affected files are being rebuilt.
      • 11:12. The affected files have been restored.

    • Residual blocks in the MPES
      2004 Mar. 12: 18:04. A user reported that residual blocks were missing from the MPES. A check showed that the procedure that copies the residual blocks into the web service had not been restarted following the upgrade. This has now been done.

    • New observation format documentation
      2004 Mar. 11: 14:20. A slightly modified version of the documentation on the new observation format has been put on the anon-ftp site. The changes from the previous version are the definition of the format for occultation records and the cometary-magnitude aperture flag.
      • The .PS and .PDF versions of this document appear to be truncated at the end of page 12. It appears that earlier versions of this document also had this problem. We are investigating.
      • 14:27. Complete 15-page .PS and .PDF versions are now on-line.

    • NEOCP start date
      2004 Mar. 9: 10:27. A user reported a problem with selecting an offset of -5 hours in the NEOCP, specifically that the entered offset was ignored. This problem was tracked down to a problem with calling a recently-extended library routine. New code has been put on-line.

    • Webserver downtime
      2004 Mar. 8: 09:36. The webserver machine is about to be upgraded to the new OS version. The downtime should be about two hours.
      • The interface card has been replaced, but the CD-ROM is still not accessible. OS upgrade has been postponed. Service has been called. Upgrade will occur after CD-ROM drive has been replaced.
      • 14:30. Service is here and the CD-ROM is being replaced.
      • 15:00. CD-ROM has been replaced. Since our VMS consultant has to leave early today, the upgrade of SCULLY has been put off until tomorrow morning.
      • Mar. 9: 10:28. The upgrade of SCULLY is beginning. Downtime should be around two hours.

    • March 6 mini-batch MPC files
      2004 Mar. 6: 12:00. A user reported that the files associated with the Mar. 6 mini-batch MPC were not updated correctly. A problem was detected in one command procedure used by the updating routines, concerning case-sensitivity of filenames. The affected routine has been made case-insensitive and the files are being regenerated.
      • 18:00. The correct files should now be in place.

    • March 5 DOU MPEC
      2004 Mar. 5: 08:59. One of the upgraded machines seems to have a problem accessing the CF's webserver machine. By chance, the DOU MPEC was prepared on an execution queue running on this machine. Various bits involving copying of data to the CF's webserver failed. The majority of these failed bits have been rerun. Some problems may remain.

    • Computer Upgrades V and the web services
      2004 Mar. 4: 15:56. One of the two remaining machines has been upgraded. The other, the webserver, has a faulty CD-ROM drive--service has been called, they won't be able to get here until tomorrow. The cluster is back and working. We are attending to a number of upgrade-related issues before reactivating the AUTOACK (and related) queues.
      • 17:01. The AUTOACK (and related) queues have been restarted.
      • 17:01. SCULLY will be upgraded sometime in the next few days.
      • 18:25. The web services have been reactivated. Please report any problems to the usual place.
      • Mar. 5: 10:00. Service arrived, will not replace SCSI card controlling CD-ROM as it is not one of theirs. We are obtaining a replacement interface card and will install it on Monday. Assuming that this fixes the problem of the non-functioning CD-ROM drive, the upgrade of the OS on SCULLY will then follow immediately.
      • Mar. 6: 10:00. Replacement interface card has been delivered overnight FedEx. Upgrade will occur on the morning on Mar. 8.

    • Computer Upgrades IV
      2004 Mar. 4: 08:56. The upgrades of the final two machines, including the web server SCULLY, in the original cluster is beginning. The web services will probably be unavailable until both machines are upgraded, probably around noon.

    • MPES oddities
      2004 Mar. 4: 08:56. Some users have reported oddities with obtaining ephemerides of numbered minor planets in the MPES. These problems are related to the ongoing computer upheaval. When the final machines have been upgraded, we will troubleshoot this problem.

    • AUTOACK and MPES II
      2004 Mar. 3: 17:47. We've finished upgrading three of our cluster machines. We have restored internal and external (web) services. Two more machines remain to be upgraded: this will occur tomorrow morning. One of the remaining machines is the web server, SCULLY. Some further disruption is expected to occur at this time.
      • As noted below, tonight's DOU MPEC has been submitted to an execution queue.

    • AUTOACK and MPES I
      2004 Mar. 3: 11:35. We've just discovered that in our attempt to standardize the list of authorized users on all machines in the cluster that we inadvertently erased five special accounts on SCULLY that are needed for, amongst other things, AUTOACK to function! We are currently adding these special accounts back into the list of authorized users.
      • The erroneously-deleted users have been reinstated.
      • However, it appears that it will be very difficult to maintain AUTOACK in a running state while the upgrades are proceeding. We are thus compressing the schedule for the upgrades in order to complete them as soon as possible.
      • The AUTOACK procedure will accept new incoming observation batches but will not ACK them automatically until the upgrades are complete. We will invoke the ACK mechanism manually at odd intervals until this time.
      • The web services are likely to be intermittent until the upgrades are complete, presumably some time tomorrow. Please be patient.

    • March 3 DOU MPEC
      2004 Mar. 2: 22:45. Due to e-mail problems mentioned below, we are forced to cancel the issuance of tonight's DOU MPEC. We hope to return to normal service tomorrow night.
      • Mar. 3: 12:05. We may not be able to resume DOU MPEC production until the upgrades are completed.
      • Mar. 3: 17:50. Following the completion of today's upgrades a decision was made to submit tonight's DOU MPEC preparation job into the execution queue, in the expectation that it will complete successfully.

    • Computer Upgrades III
      2004 Mar. 2: 09:16. The upgrades of the machines in the original cluster to the new version of the OS begin today. The webserver machine, SCULLY, is currently scheduled for the upgrade in the afternoon of March 3. The web cgi scripts that run on SCULLY will be unavailable during the upgrade of this machine. The scripts may also have problems if they need data off a disk that is being served to the cluster by another machine that is being upgraded.
      • Mar. 2: 11:56. The upgrade of SENECA is about to begin.
      • Mar. 2: 14:00. The upgrade of SENECA has finished. Some e-mail problems with the new setup have emerged. We are working on fixing them.
      • Mar. 2: 22:45. The e-mail problems persist. It appears we need to install a patch to fix them.
      • Mar. 3: 09:30. Reexamining the e-mail problem. It appears to affect only the user whose home directory resides on a disk attached to SENECA. Other users e-mailing from SENECA are apparently unaffected.
      • Mar. 3: 10:13. The e-mail problem has been tracked down and fixed. E-mail destined for the affected user is now being delivered.

    • Minor Planet Ephemeris Service
      2004 Feb. 29: 18:10. A user asked if it would be possible to include a listing of all designations associated with a designation that was entered into the MPES. The MPES already indicates the principal designation if a non-principal designation was entered. We are of the opinion that it is not appropriate to list all the non-principal designations on the pages returned by the MPES. However, we have put in links to the interactive minor-planet designation script to make it easy to obtain the necessary information.

    • Minor Planet Designations script
      2004 Feb. 29: 11:20. A user reported problems with the interactive minor-planet designation script. An investigation showed that the data files for this script were still being prepared on the old VAXes, where the necessary files are no longer being maintained. The procedure has now been ported over to the cluster and the necessary data files have been rebuilt. One minor problem remains: the returned page says "You requested information on the designation" followed by two unprintable characters, rather than informing you which designation you were searching for.
      • 11:35. While attempting to fix the last problem, the old Unix C compiler's "oops, compilation errors, let's create a blank executable rather than leave the old executable intact" gotcha has bitten again. Users will receive an error message if they try to access the script.
      • 11:40. A new executable that fixes the problem about displaying the designation being search for has been installed.

    • Timing of DOU MPECs
      2004 Feb. 25: 12:00. While testing of the new queuing system on the new cluster is being undertaken, the timing of the issuing of DOU MPECs may be unpredictable (e.g., MPEC 2004-D35 and MPEC 2004-D36). Regrettably, it is not possible to indicate in advance which future DOU MPECs will be affected and by how much. Once the testing of the new queues is completed, the DOU MPEC will revert to the normal schedule.

    • NEA+*.DAT files
      2004 Feb. 23: 14:30. A user reported that the NEA+*.DAT odd-epoch files on the MPCORB ftp page (and on the mirror sites) appear to be missing objects with designations assigned after mid 2003. This is an error and we are investigating.
      • Feb. 23: 14:45. Investigation shows that the procedure that builds those files depends on some files that have not been prepared since the MPES underwent its major upgrade. We are rewriting the procedure.
      • Feb. 23: 16:20. The procedure to build the NEA+*.DAT files has been recoded and has been extended to prepare the corresponding files for dates from -10 to +14 days of the current date. The files on the local MPCORB site have been updated. The mirror sites may not get updated until tomorrow.

    • DOU MPEC
      2004 Feb. 12: 09:50. The overnight DOU MPEC batch job stalled. We aborted the procedure and examined the log file. We then ran the procedure interactively to try and reproduce the problem that stalled the batch job. No problem was detected and the circular was prepared (with a new MPEC number) without incident. A number of NEO observations that should have been included on the circular were inadvertently omitted, they will appear on the next DOU MPEC.

    • Computer Problems II
      2004 Feb. 11: 09:50. Two disks attached to the cluster member with the dodgy CD-ROM drive are unavailable to the other cluster members. This has caused some problems with updating elements sets for the Web cgi scripts. The two disks are in an external box: we will reattach them to another cluster member and rerun the overnight procedures that failed.
      • 10:44. The external disk box has been attached to another machine and the two overnight procedures that failed have been rerun.

    • Computer Problems
      2004 Feb. 10: 12:00. While replacing the CD-ROM drive on one of our cluster machines, a cluster problem developed which is "stiffing" the other cluster machines. We are attempting to fix the problem.
      • 12:30. A number of cluster-shared disks are refusing to dismount cleanly. One disk is refusing to be shared among the cluster members. We are trying to understand why this is the case. The Web server machine will probably be unavailable for a number of hours.
      • 17:00. The cluster was returned to normal operation after much investigation. Changes planned for the near future will reduce the likelihood of a similar outage in the future. Web server should now be operating normally.

    • Files for the ECS
      2004 Feb. 8: 18:00. A hardware problem on one of our machines is preventing us putting up the monthly files for the ECS. The files will be available as soon as possible after the machine is repaired.
      • Feb. 9: 09:50. We have called for service.
      • Feb. 9: 10:50. Since the machine is rather old, the required motherboard is not available at the local service office. It will have to be flown in from Kentucky. Service should be here tomorrow.
      • Feb. 10: 11.00. Motherboard did not make it in time. Service is here for an unrelated problem on one of our cluster machines that requires replacement of a CD-ROM drive. Motherboard should be in tomorrow afternoon.
      • Feb. 11: 15.30. The motherboard has been replaced. The posting of the files in the telnet and Web ECS is underway.

    • Special Monthly MPECs
      2004 Feb. 5: 22:40. A user pointed out that the date of last observation of (25143) was incorrect on MPEC 2004-C20. Investigation showed that during the filing associated with the 2004 Feb. 7 batch of MPCs the new numbered observations were not added to the files of previously-published observations. This is now being done.
      • The last dates of observation for the numbered objects in the special MPECs are suspect. Correct dates may be obtained through the MPES.
      • We will modify some filing procedures to ensure that this does not happen again.

    • NEOCP
      2004 Feb. 4: 18:30. The list of previous NEOCP objects has been moved to a separate page. The list typically has 100 objects listed at any one time and by moving the list to another page the size of the NEOCP pages has been halved.

    • MPCs in ADS
      2004 Feb. 4: 12:30. The MPECs and IAUCs have been indexed in the Astrophysics Data System for a number of years. The MPCs are now also included. At the present time, indexing is available for circulars issued since June 2003. Future issues will be included as a matter of routine. Other back issues will be included as time permits.

    • NEOCP uncertainty maps II
      2004 Feb. 4: 11:10. The uncertainty maps in the NEOCP have been further extended to indicate those variant orbits that will pass within 0.0027 AU (roughly one lunar distance) sometime over the next 100 hours. Such orbits will be flagged with black triangles on the uncertainty maps and by "***" on the lists of offset coordinates.
      • Feb. 4: 11:30. Counts of the number of variant orbits in each of the four flagged categories are now displayed to the right of the uncertainty plot.

    • "Denial of service" use of web script
      2004 Feb. 3: 12:00. It has been noted that a single user has been sending many single requests (in one instance, seven requests in one second) to the supernova checker script, rather than using the batch facility. Access to the script (and to the mp checker script) has been disabled for the IP address in question (and will be expanded to the domain if the user roves in IP space) and the user will see a message explaining why access is blocked. This is probably the reason for the non-responsive NEOCP reported by a user last night.

    • Machine upgrade
      2004 Feb. 3: 11:00. The CF box that acts as our gateway to the CF's web server machines has apparently been up for 200 days and is urgent need of OS patches. The CF has already started installing the required patches and the work is expected to take 50 minutes. Until this work is completed some pages on the MPC's website may not get updated.

    • NEOCP uncertainty maps
      2004 Jan. 29: 10:10. The uncertainty maps in the NEOCP have been extended to include an indication of objects that are close to the earth. Three colors are used to indicate objects at various distances:
      • green, for objects that are more then 0.05 AU from the earth at the time used for the uncertainty map;
      • orange (supposedly, the color appears to depend on your browser/OS) for objects between 0.05 and 0.01 AU from the earth
      • red, for objects within 0.01 AU.
      On the list of offset coordinates, variant ephemeris positions that are orange on the map are flagged by "!" and positions that are red on the map are flagged by "!!".

      The standard color for non-NEOCP uncertainty maps has also been changed from red to green.

      Jan. 29: 13:40. Following feedback from an observer, the shade of green used in the uncertainty maps was darkened, in order to distinguish it from the white background. In addition, another observer reported a problem with variant ephemerides offset from the current date: this has been fixed.

    • Computer Upgrades II
      2004 Jan. 27: 14:30. The MPC is planning to merge the existing cluster machines with the new cluster machines on Feb. 6. Following the merge, the existing cluster machines will have their operating systems upgraded to the latest version. When the web-server SCULLY is upgraded, it will be off-line during the upgrade. At the present time, this is expected to occur on Feb. 9. Further information will be posted closer to the date.

      On a related matter, we are pleased to report significant improvements in a number of core processes used by the MPC (an example being a decrease from 260s to 100s in the time taken to do a certain set of tasks). Further performance improvements (over and above the basic 3x increase in raw CPU cycles available) are expected once the two clusters are merged.

    • Non-standard-compliant browser behavior
      2004 Jan. 24: 10:00. A user reported that the text area in the New Object Ephemeris Generator was not wide enough to enter an 80-column observational record when using MS IE in conjunction with Windows XP. The text area has been extended to 81 columns, but HTML text areas are defined in terms of characters so an inability to accept the specified number of characters is non-standard behavior. Users should report this to MS and request a fix.

    • No residuals blocks for numbered objects in MPES
      2004 Jan. 22: 10:00. A user reported that residual blocks for numbered objects were not available. The reboot of the webserver two days ago caused some logical disknames to not be set up correctly (various configuration work on the existing cluster is being done in order to facilitate moving those machines into the new cluster). This anomaly has been corrected.

    • Overheating office
      2004 Jan. 20: 10:00. For some as-yet-unexplained reason, the heat in the office containing the webserver and the new cluster machines was turned up yesterday. The resulting desert-like heat affected the computers and caused at least one of the machine to reboot. Temperatures have now been returned to normal, but short-term replacement of the heating/cooling unit is a priority.

    • NEOCP variant ephemerides
      2004 Jan. 14: 16:00. Following a request from a user, the NEOCP has been extended to allow access to ephemerides of variant orbits used in preparing the uncertainty plots. The links to the ephemerides of these variant orbits may be found on the page listing the offsets from the nominal solution.
      • Jan. 15: 16:15. The same user requested that the variant ephemerides be numbered on the offsets page. This has been done.

    • Web space quota exceeded?
      2004 Jan. 3: 16:30. There appears to be a problem with the disk quota on the Computation Facility's boxes that may prevent from us from updating webpages and datafiles. Unfortunately, the available commands for determining disk quota and disk-space usage are giving conflicting information, so the source of the quota problem is unknown. We are seeking urgent clarification from the CF and are awaiting a response.
      • 16:40. At least one other MPC staff member has reported problems with their CF account. No response yet from the CF.
      • 20:55. The problems we saw seem to have been cleared up, but there has been no explanation from the CF as to the cause of the problems.

    • Links to 2004 MPECs
      2004 Jan. 1: 21:10. A problem arose with the HTML links between MPECs issued in 2004. This was caused by the inherent lack of intelligent error handling in the ftp protocol, triggered at the start of the new year by a file that was not copied from the VAX to the Alpha cluster when the MPEC-filing routines were ported. The missing file has been copied, the affected procedure has been recoded and future circulars should not suffer from this problem.