OK I'll add that (what IW does on setting an OOME) to the javadocs.
Mike
Jed Wesley-Smith wrote:
Mike,
regarding this paragraph:
"To workaround this, on catching an OOME on any of IndexWriter's
methods, you should 1) forcibly remove the write lock
(IndexWriter.unlock static method) and then 2) not call any methods on
the old writer. Even if the old writer has concurrent merges running,
they will refuse to commit on seeing that an OOM had occurred."
I'm not sure that an IndexWriter is particularly useful once its
hitOOM flag is set to true, whether autocommit is true or not. Once
it is true you can't do anything with it (it reverts to its last
commit point and stays there) and need to discard it. I am
suggesting that this could be documented as it is not immediately
obvious without coming across it and debugging it. That being said,
the VM is probably not that useful once OOMEs are flying around
anyway :-)
cheers,
jed.
Michael McCandless wrote:
Jed Wesley-Smith wrote:
Yeah, I saw the change to flush(). Trying to work out the correct
strategy for our IndexWriter handling now. We probably should not
be using autocommit for our writers.
autoCommit=true is deprecated as of 2.4.0, and will go away when we
finally get to 3.0, so I think switching to false, and possibly
changing your app to periodically commit() if you were relying on
those semantics, is a good step forward.
It was brought up by others that the OutOfMemoryError handling
requirements are a fairly strong part of the contract now - but
aren't documented. Do you think the last paragraph below should be
incorporated into the class JavaDoc?
Well, that paragraph is a workaround for the issue you hit, which
only applies when autoCommit is true, so going forward (or, if you
use autoCommit=false) you should simply close the IndexWriter.
---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
