We are using TRIM 184.108.40.2060 and we have been having issues with document content indexing for a while. After the implementation DCI worked without any issues for a month or so and then started failing repeatedly. Our users will report that they get an error when performing a content search, I will then delete the TRIMPending.bin fie and restart the event service to fix it. Since it started happening three or four times every day, we requested the assistance of our TRIM vendor and they recommended to move the event service to another server. So another server was configured to host the TRIM event service (which is a virtual machine cloned from the actual production server) and we did a complete re-index of the document store on the new server (which took 52 hrs to complete and thousands of documents were added to the quarantine). DCI worked fine on the new server for a couple of weeks and it started to fail again (three to four times a day as before). The workgroup server log, trimdci log and windows application event log were looked at by TRIM vendor and HP, but could not identify the exact cause of the DCI failure. There are a lot of password protected documents and some MS Word documents which TRIM can not open and read in the quarantined area. Our TRIM vendor recons that the quarantined word documents are actually RTF documents with the extensions somehow changed to .doc and TRIM seems to have an issue with handling these documents. Could these documents be causing the DCI failure? This issue has been going on for months now and we have not been able to figure out why.
The workgroup server is Windows 2003 Enterprise SP2 and all TRIM Context clients run on Windows XP Professional SP2 and Windows 7 Professional. Could anyone suggest why this could happening or what can be done to identify the cause of the failure? Any help would be highly appreciated.
Hey, Usually when the DCI starts failing 3 or 4 times a day, i have found that index is curropted someway. What I do is make a new chain and make the old train read only seems to work for about 2 - 3 months before it fails again.
Maybe putting the RTF file type in the TES and making it bypassed for DCI files might help. Are you getting any errors in the log files or application eventvwr logs..
Thank you very much for your response. I already did what you suggested (to create a new folder in the index chain) when DCI kept on failing last time and DCI seems to work fine after that. Below is taken from the TRIMDCI log file when the first failure (in the latest sequence of failures) occurred. As instructed in the log, we contacted HP (through our TRIM vendor) and gave the the error code 'HwDelete-03'. But we did not recieve any useful information from HP. Have you or anyone else in the forum has seen this error?
<<DCI LOG - START>>
05:55:49 0 XX Content Index Engine message, Removing references to changed and deleted documents. 05:55:49 0 XX Content Index Engine message, (including 4,398 deferred deletions). 05:55:49 0 XX Content Index Engine message, (Deleted 7) TRIMDB_XX_1037680 2,806 words. 05:55:49 0 XX Content Index Engine message, (Deleted 50) TRIMDB_XX_1037737 174 words. ............................... ............................... 05:56:07 0 XX Content Index Engine message, (Deleted 27538) TRIMDB_XX_1058289 33 words. 05:56:07 0 XX Content Index Engine message, TRIMDB_XX_1056201 720 words. 05:56:07 0 XX Content Index Engine message, (Deleted 27572) TRIMDB_XX_1058975 912 words. 05:56:10 0 XX ***** A unrecoverable error occurred with the content indexing engine, 26/10/2010 3:56:10 PM - An Internal Error has occurred. If this reoccurs, contact vendor and quote code: HwDelete-03.
I know why HP helpdesk couldnt give you a direct answer ont he HwDelete error message as it is an isys error message and not a TRIM error. I suggest you could email ISYS team but they probably wont be able to help as TRIM uses a cut down version of isys. Maybe HP can contact them for you if you push it. Alternatively if you can find one of the documents which courses the engine to break and forward it to HP, they maybe able to do something with it... (Big Maybe).
What Ive read on a simlar issue is that it points towards a curropt file being processed (as you have stated it looks like your .doc's are seen as RTF documents).
Do you know if the support desk spoke with the Isys team ?
As you mentioned, HP responded to our vendor that HwDelete-03 is an ISYS error. I was told by the vendor that thay are going to escalate the issue with HP. I have not heard anything from them after that.
From the section of the DCI log in my above posting, is it possible to identify the document that might be causing the DCI to fail? it looks like the DCI fails when it tries to delete removed references, not when it tries to index a document. If that is correct, what does that indeicate? Could you kindly advise?