JChem Index issues

User c0e481a82c

24-08-2007 09:21:25

Hi,





I've recently come across a problem which has caused a couple of days downtime. Essentially, it appeared as though the index had become corrupted. Substructure searches would not finish after well over 12 hours, yet there was no error message. Restarting Tomcat and Oracle made no difference. I figured the issue might be concerned with the index so rebuilt the index, only to be told immediately by Oracle that the tablespace on which the indexes are built was full. Unfortunately, we're without a DBA at the moment so I'm unable to give you more information about the issue. Therefore, I would just like to let you know that this happened, and I'm puzzled as to why JChem was unaware that Oracle had an issue. The solution was obvious once an error message was thrown, but why didn't JChem throw it sooner?





The JChem version is 3.2.6.





Regards,





Phil.

ChemAxon aa7c50abf8

24-08-2007 12:29:39

Phil,





Thank you for reporting this problem.





Has adding new files to the table space solved the search problems?
Quote:
why JChem was unaware that Oracle had an issue.
If Oracle raised an exception, I expect JChem Cartridge to have passed it further on. At this point, I have no reason to assume that the JChem code involved in structure search swallowed any error message coming from Oracle. What I find much more likely is that Oracle suspended operations on the tablespace and was awaiting external intervention to solve the shortage problem.





Another relevant question might be: what role may JChem Cartridge have played in the unexpected tablespace exhaustion? Was there an unepxected burst in the growth of table space usage attributable to JChem Cartridge? If there was, was this burst inconsistent with the size of the jobs JChem Cartridge has been given?





Thanks


Peter