- Package:
- libsqlite3-0
- Source:
- sqlite3
- Description:
- SQLite 3 shared library
- Submitter:
- Christoph Egger
- Date:
- 2026-11-01 10:03:02 UTC
- Severity:
- important
Hi! This is the second time sqlite managed to corrupt its database in the last week on this system. The sqlite is located on a ext3 filesystem with lots of free space. Unfortunately the corrupted sqlite3 contains private keys so I can't provide the database or one of the daily backups. Christoph # sqlite3 asdf1234 vacuum Error: database disk image is malformed
Hi Cristoph, May you give me pointers? Is it a big database, uses concurrent access, is the CPU load high? Any chance you may reproduce it with a non-confidental dataset? I think the best would be to subscribe to the SQLite3 users list[1]; upstream may help without giving out your data. Regards, Laszlo/GCS [1] http://mailinglists.sqlite.org/cgi-bin/mailman/listinfo/sqlite-users
Hi! "László Böszörményi (GCS)" <gcs@debian.org> writes: ~1.3 MiB There's two processes accessing the database. It also currently runns in WAL mode. CPU can get max'ed. This is just a 900MHz mipsel and it's some things that require quite some computation I can try. Not in the next 2 weeks though, sorry. Christoph
In the meantime, another serious sqlite corruption bug has been found for DBs running in wal mode, which I have been reported against Ubuntu: https://bugs.launchpad.net/ubuntu/+source/sqlite3/+bug/2171337. Maybe it should get fixed in Debian as well. On Fri, 10 Jun 2016 11:58:06 +0200 Christoph Egger <christoph@christoph-egger.org> wrote: > Hi! > > "László Böszörményi (GCS)" <gcs@debian.org> writes: > > On Thu, Jun 9, 2016 at 4:55 PM, Christoph Egger <christoph@debian.org> wrote: > >> This is the second time sqlite managed to corrupt its database in > >> the last week on this system. The sqlite is located on a ext3 > >> filesystem with lots of free space. Unfortunately the corrupted > >> sqlite3 contains private keys so I can't provide the database or one > >> of the daily backups. > > > May you give me pointers? > > Is it a big database, > > ~1.3 MiB > > > uses concurrent access, > > There's two processes accessing the database. It also currently runns in > WAL mode. > > > is the CPU load high? > > CPU can get max'ed. This is just a 900MHz mipsel and it's some things > that require quite some computation > > > Any chance you may reproduce it with a non-confidental dataset? > > I can try. Not in the next 2 weeks though, sorry. > > Christoph > > -- > 9FED 5C6C E206 B70A 5857 70CA 9655 22B9 D49A E731 > Debian Developer | Lisp Hacker | CaCert Assurer > >
Hi Matthias, Thanks for the heads-up. Is there any test case to check if the bug is fixed or not? It seems the fix is intrusive, a full package backport would be better. Does it affect samba in any way? Regards, Laszlo/GCS