RE: [Monotone-devel] how to recover damaged database?

[Kelly F. Hickel]

<snip>

> > What would be helpful here would be if you could reproduce the
> problem
> > while passing --debug or --dump?  --debug generates copious
> debugging
> > output to stderr; --dump takes a file argument, and when an error
> like
> > the one you see here occurs it dumps (slightly less) copious
> debugging
> > output to that file.  Either would probably be sufficient.
> >
> > (Well, I guess it would also be useful if you could make the
> offending
> > CVS repo available, but I'm guessing from the branch names that that
> > isn't going to be possible :-).  You might want to look over the
> > debug output as well to double-check there's nothing sensitive
> there,
> > but I'm guessing it should be mostly okay, filenames are probably
> the
> > most sensitive thing that might be there.)
> 
> OK, I have the log, it's 23MB unzipped, 2.3MB zipped.  Probably too
> big
> to send to the list.  Should I send it to someone via direct email (if
> so, who)?
> 
> >
> > Thanks,
<snip>

In case it's useful, here's the tail end of the debug log, from a bit
before the first occurrence of the word "violated".

-Kelly

mtn: binding 0 parameters for COMMIT
mtn: binding 0 parameters for BEGIN EXCLUSIVE
mtn: trunk has 1038 entries
mtn: examining next commit [t:1046918479]
[p:qpea_replay/create_replay.pl] [a:kfh] [c:file create_replay.pl was
initially added on branch B_Maint_3_0_1.
mtn: ]
mtn: last modification time is 0
mtn: building new cluster [t:1046918479] [a:0] [c:176]
mtn: examining next commit [t:1047008539]
[p:qpea_replay/qpea_replay.dsp] [a:kfh] [c:file qpea_replay.dsp was
initially added on branch B_Maint_3_0_1.
mtn: ]
mtn: expiring cluster
mtn: last modification time is 0
mtn: building new cluster [t:1047008539] [a:0] [c:131]
mtn: examining next commit [t:1047008539] [p:qpea_replay/qpea_replay.c]
[a:kfh] [c:file qpea_replay.c was initially added on branch
B_Maint_3_0_1.
mtn: ]
mtn: examining cluster 0 to see if it touched 766
mtn: last modification time is 0
mtn: building new cluster [t:1047008539] [a:0] [c:154]
mtn: examining next commit [t:1047008539]
[p:qpea_replay/qpea_replay.dsw] [a:kfh] [c:file qpea_replay.dsw was
initially added on branch B_Maint_3_0_1.
mtn: ]
mtn: examining cluster 0 to see if it touched 768
mtn: last modification time is 0
mtn: building new cluster [t:1047008539] [a:0] [c:175]
mtn: examining next commit [t:1047188402] [p:qpea_replay/makefile]
[a:kfh] [c:file makefile was initially added on branch B_Maint_3_0_1.
mtn: ]
mtn: expiring cluster
mtn: revision.cc:50: invariant 'I(null_id(edge_old_revision(i)))'
violated
mtn: saving current work set: 3 items
mtn: finished saving work set
mtn: contents of work set:
mtn: Current work set: 3 items
mtn: ----- begin 'cmdline_string' (in int cpp_main(int, char**), at
monotone.cc:274)
mtn: 'mtn', '--debug', '--db=eadev_import.mtn',
'--branch=com.mqsoftware.qpasa.qa', 'cvs_import',
'/usr/cvsroot/master/qa'
mtn: -----   end 'cmdline_string' (in int cpp_main(int, char**), at
monotone.cc:274)
mtn: ----- begin 'locale_string' (in int cpp_main(int, char**), at
monotone.cc:278)
mtn: en_US.UTF-8
mtn: -----   end 'locale_string' (in int cpp_main(int, char**), at
monotone.cc:278)
mtn: ----- begin 'full_version_string' (in int cpp_main(int, char**), at
monotone.cc:283)
mtn: monotone 0.26 (base revision:
4342565107f26ceda955b66c66b5b7ec152f314e)
mtn: Running on          : Linux 2.6.9-34.ELsmp #1 SMP Wed Mar 8
00:27:03 CST 2006 i686
mtn: C++ compiler        : GNU C++ version 3.4.5 20051201 (Red Hat
3.4.5-2)
mtn: C++ standard library: GNU libstdc++ version 20051201
mtn: Boost version       : 1_32
mtn: Changes since base revision:
mtn: unknown
mtn: -----   end 'full_version_string' (in int cpp_main(int, char**), at
monotone.cc:283)
mtn: prepared statement ROLLBACK
mtn: binding 0 parameters for ROLLBACK
mtn: statement cache statistics
mtn: prepared 32 statements
mtn: 47 executions of BEGIN EXCLUSIVE
mtn: 46 executions of COMMIT
mtn: 2 executions of DELETE FROM file_deltas WHERE id = ?
mtn: 4 executions of DELETE FROM files WHERE id = ?
mtn: 56 executions of DELETE FROM rosters WHERE id = ?
mtn: 187 executions of INSERT INTO file_deltas VALUES(?, ?, ?)
mtn: 705 executions of INSERT INTO files VALUES(?, ?)
mtn: 72 executions of INSERT INTO revision_ancestry VALUES(?, ?)
mtn: 288 executions of INSERT INTO revision_certs VALUES(?, ?, ?, ?, ?,
?)
mtn: 72 executions of INSERT INTO revisions VALUES(?, ?)
mtn: 56 executions of INSERT INTO roster_deltas VALUES(?, ?, ?)
mtn: 71 executions of INSERT INTO rosters VALUES(?, ?)
mtn: 72 executions of INSERT into revision_roster VALUES (?, ?)
mtn: 1 executions of ROLLBACK
mtn: 1 executions of SELECT 1 FROM revisions LIMIT 1
mtn: 1 executions of SELECT 1 FROM rosters LIMIT 1
mtn: 1 executions of SELECT base FROM file_deltas WHERE id = ?
mtn: 3 executions of SELECT data FROM files WHERE id = ?
mtn: 72 executions of SELECT data FROM revisions WHERE id = ?
mtn: 54 executions of SELECT data FROM rosters WHERE id = ?
mtn: 1 executions of SELECT delta FROM file_deltas WHERE id = ? AND base
= ?
mtn: 1162 executions of SELECT id FROM file_deltas WHERE id = ?
mtn: 1257 executions of SELECT id FROM files WHERE id = ?
mtn: 288 executions of SELECT id FROM public_keys WHERE id = ?
mtn: 288 executions of SELECT id FROM revision_certs WHERE id = ? AND
name = ? AND value = ? AND keypair = ? AND signature = ?
mtn: 463 executions of SELECT id FROM revisions WHERE id = ?
mtn: 71 executions of SELECT id FROM roster_deltas WHERE id = ?
mtn: 182 executions of SELECT id FROM rosters WHERE id = ?
mtn: 7201 executions of SELECT node FROM next_roster_node_number
mtn: 71 executions of SELECT parent FROM revision_ancestry WHERE child =
?
mtn: 56 executions of SELECT roster_id FROM revision_roster WHERE rev_id
= ? 
mtn: 7201 executions of UPDATE next_roster_node_number SET node = ?
mtn: fatal: std::logic_error: revision.cc:50: invariant
'I(null_id(edge_old_revision(i)))' violated
mtn: 
mtn: this is almost certainly a bug in monotone.
mtn: please send this error message, the output of 'mtn --full-version',
mtn: and a description of what you were doing to
address@hidden
mtn: discarding debug log, because I have nowhere to write it
mtn: (maybe you want --debug or --dump?)