While not a common issue, there are times when, for some reason or another, a host may drop a connection during an indexing operation. If this occurs during a normal (initial) index operation, Sphider records this and, upon re-connection, allows the indexing operation to resume.
If the disconnect occurs during a RE-index, however, Sphider would start the next re-index operation at the beginning.
In very rare instances, an unstable host might prevent a total re-index from ever completing.
The latest releases (Sphider 4.0.0-MB and SphiderLite 2.0.0) are able to allow a re-index to complete, under the right circumstances.
IF you find the host you are dealing with is prone to frequent disconnects, following these steps may prove beneficial:
1) Before attempting the re-index operation, check the Clean Tables tab. If temp and/or temp2 show non-zero, clean them. This is ONLY during the first re-index run. If the re-index run ends with the "completed" notice, you are finished. There was no interruption.
2) If you did not received a "completed" notice, the run was interrupted by a disconnect. BEFORE logging off or doing other indexing operations, you will note them temp2 (and possiibly temp) are NOT empty. DO NOT CLEAN THEM! Do another re-index run. Sphider will attempt to resume the re-index at the point it was interrupted. A couple pages may be processed again before Sphider skips ahead to where it was interrupted. Keep repeating this process until you get the "completed" notice. At this point, both temp and temp2 should be empty.
Note that if the session is interrupted (you log off, close your browser, or perform operations on other websites, the process will be damaged and any subsequent re-index of the affected site will begin anew.
1 post • Page 1 of 1