Me again. Continued testing and think I found a more useful error message that helped me to solve the problem.
2008-01-13 08:27:32 (4001d4a0): Found /mnt/array1/Music/Album Artwork/Download/8DC1DD60EC7C97F3/09/00/10
2008-01-13 08:27:32 (4001d4a0): Found 10.. recursing
2008-01-13 08:27:32 (4001d4a0): Found /mnt/array1/Music/Album Artwork/Download/8DC1DD60EC7C97F3/09/00/10/8DC1DD60EC7C97F3-2F960DF25E02008-01-13 08:59:07 (4001d4a0): select(8, 5.000000)
2008-01-13 08:59:08 (4001d4a0): Query: commit transaction
2008-01-13 08:59:08 (4001d4a0): Error: database is full
2008-01-13 08:59:08: Aborting
2008-01-13 08:59:08 (4001d4a0): Processing rendezvous message
2008-01-13 08:59:08 (4001d4a0): Rendezvous socket closed (daap server crashed?) Aborting.
2008-01-13 08:59:08: Aborting
The database is full message made me do a quick df. It turns out that both the cache and log folders are on the TSL’s small / partition. With a large database like mine (plus running at debug 9) I was maxing out the partition, causing the write error and the crash.
The TSL comes with a nasty mt-daapd pre-installed and it keeps its files in a pre-created directory on the main disk array. I’ve changed the conf to put the log and db there and now it’s running perfectly.
I’ll go back to the original instruction writers and provide them with the update.
Cheers and good luck with the db rebuild!
M