[Lurker-users] Lurker Database Corruption
Brought to you by:
terpstra
|
From: Jamin W. C. <jco...@as...> - 2002-09-20 15:54:00
|
It appears that Lurker frequently corrupts it's database if Lurker is improperly stopped (or not stopped and the system is restarted). I realize that both of these are not normal desired operation. I'm just wondering if it wouldn't be possible to incoporate more resilience with regard to Lurker's database. My particular situation I'm currently experiencing is due to sporadic system reboots (I may have finally located the cause). The system recovers very well from these reboots (thanks to ext3), with the single exception of Lurker. On some occassions Lurker restarts and appears to work (provides listing of message headers) but is unable to retrieve any message bodys. On others, Lurker fails to start at all. The following are the error messages I've seen in my system logs for this: lurkerd[728]: CORRUPT dataset: Imported a message from a new mbox which has timestamp earlier than present. This means an mbox was added to the config file after the existing database had passed this timestamp. All timestamps in this mbox will be wrecked. lurkerd[2300]: Program Fault (signal 6) - attempting to salvage db In all cases stopping lurker, purging the database, and restarting lurker correct the situation. I'm using version 0.1e. -- Jamin W. Collins |