A customer received an assertion error 200505 checksum failure on page xxx on a server version 126.96.36.19952.
The database is an approximately 300 GB remote database using SQL Remote replication, there was no full database backup.
I successfully unloaded the database using dbunload without ordering the data and therefore can recover the database using the unload, reload and reset of the offsets as explained in the documentation. However the unload and reload will take approximately 48 hours and I don't want to be down for that long.
My idea was to do the unload, reload, and offset thing and then apply the transaction logs from the online version of the corrupt database however SQL Anywhere is smarter than I and wouldn't apply the files saying they belong to a different database. By the way it would be nice if you had an option that would allow this.
I believe I can translate and apply the log files on a daily basis and then reset the offset to the ending offset of the last translated log file, (please verify). However I would rather fix the problem associated with the bad checksum since I assume the successful unload indicates the error is in an index page.
I determined which table is associated with the page with the bad checksum. Is my best approach to reorganize this table or drop the primary and foreign kay as well as any other indexes and rebuild them.
Thanks in advance
asked 22 Jun '15, 16:07