Happy New Year! This is first time that I visiting here after New Year. My
client uses eXist1.4.0 and database is corruppted due to power outage. They
followed recovery procedure and at the end they started re-indexing.
Re-indexing never finished probably even after 20 hours. They have large
database. Could Re-indexing take this long? How do we know if re-indexing is
still going on or process is hung? If re-indexing never end, how do they
restore database if they don't have backup?
View this message in context: http://exist.2174344.n4.nabble.com/How-long-re-indexing-could-take-tp4328357p4328357.html
Sent from the exist-open mailing list archive at Nabble.com.
From: Wolfgang Meier <wolfgang@ex...> - 2012-01-25 21:17:39
> They followed recovery procedure and at the end they started re-indexing.
Normally if eXist detects a crash it should start a recovery run on
its own and reindex afterwards. So I guess this did not work?
> Re-indexing never finished probably even after 20 hours. They have large
> database. Could Re-indexing take this long?
It will certainly take some time since all content in the db is read.
20 hours seems a bit long. If the reindex finished, you should see
something in the log files.
> How do we know if re-indexing is
> still going on or process is hung?
You could trigger a java heap dump and see if there are active threads
working in index-related classes. Also, the logs should contain some
information which indicates if processes are still operating.
> If re-indexing never end, how do they
> restore database if they don't have backup?
Kill the db but do not start it immediately again! Go to the data
directory and remove all .log files in there. Next I would recommend
to launch the ExportGUI and create an export first.
Afterwards, apply the complete reindex procedure: remove all files
from the data directory except dom.dbx, symbols.dbx and
collections.dbx and reindex.
Get latest updates about Open Source Projects, Conferences and News.