Running latest beta- Windows 7 Home Prem.
4478.LogosCrash.txt
1581.LogosError.log
0830.LogosIndexerCrash.txt
7345.LogosIndexerError.log
Hi Whyndell
This says that one of your files is corrupt
Suggest you remove C:\Users\Whyndell Earl\AppData\Local\Logos4\Data\0yzboier.jcx\LibraryCatalog\catalog.db
Logos will take time on restarting to prepare your library again
Graham
Graham tried your suggestion here are the files:
8814.Logos.log
1680.LogosCrash.txt
5811.LogosError.log
This is now saying that another file is malformed -
C:\Users\Whyndell Earl\AppData\Local\Logos4\Documents\0yzboier.jcx\UserPreferences2\PreferencesManager.db
Worth checking your disk to see if there are some wider issues here.
Graham- i don't see the referenced file? I only have a Usereventmanager folder
My bad found it- why do I have two folders ?
Ok deleted that folder- restarted Logos looks like it is building my library fingers crossed
Hope it works.
Most of us do. The old one relates to an older version of Logos (the old one uses Sync v1, the new one Sync v2). I guess they kept it in case something went wrong in the upgrade.
Ok looks like that worked Logos is open and indexing- sure would like to know what caused all that. But at least I learned a fix.
Ok looks like that worked Logos is open and indexing- sure would like to know what caused all that.
That's good news
But at least I learned a fix.
Please bear in mind, that fix only addresses that specific problem[:)]
10-4
sure would like to know what caused all that.
So would Logos. Problems like this occur sufficiently regularly to suggest there is an underlying problem that needs to be fixed or worked around, but not so regularly that there's enough data to diagnose the problem.
We would very much like to know what causes this problem. Unfortunately, we have never reproduced the problem in-house, even though we have stress tests to try to ferret out problems (and SQLite itself has very thorough tests).
It's possible that there's a bug in our code that causes the problem, but it's also possible that it's caused by bad hardware (see http://www.sqlite.org/howtocorrupt.html). Microsoft has a fairly depressing paper about the reliability of typical consumer hardware: http://research.microsoft.com/pubs/144888/eurosys84-nightingale.pdf.
We are considering switching to a WAL journal, which should be more robust (http://www.sqlite.org/draft/wal.html). We haven't until now because it would have caused file-locking conflicts with anti-virus software; however, recently support for persisting the WAL was added, which should allow us to work around that issue.
We are trying to work on solutions that can be deployed to fix this problem, but as mentioned, it's not entirely clear what the problem even is.
