Well this was an EOF exception so the database file was truncated, is there an away to recover the non corrupted documents?
Simple scan through the collection until the error occurs again? I am going to work on some changes to propagate the IOException into the other methods. Todd Vadim Gritsenko wrote: > Todd Byrne wrote: > >> In getBTreeNode(long page, BTreeNode parent)(line ) if an exception gets >> thrown for what ever reason it just gets ignored and null returned. This >> seems problematic because nothing that calls this method checks for >> null. Worse seems getChildNode(int idx)(line 534) doesn't check and is >> called in about 11 places and none of their callers check for null. >> >> My hunch is to let the exception be thrown and then we will get meaning >> full stack traces instead of NPEs later and have to search the log files >> later. If an exception was thrown like PageNotFound we could catch it in >> the result code and gracefully skip the document. >> >> Thoughts Ideas? > > > The only exception which should be thrown from this method is > IOException, as far as I can see. And I agree, I think it is better to > get IOException rather then rather pointless NPE. > > So for a quick fix, I'd try and change method signature(s) to include > IOException and pass it up instead of eating it. > > > It brings though a point of how xindice should handle exceptional > situations like this. If it is at all possible, it should recover. E.g., > it is better to be able to extract some documents from damaged > collection then none at all. > > > Vadim
