Re: Bug: fatal error with notmuch new, second run starts indexing all over again

Subject: Re: Bug: fatal error with notmuch new, second run starts indexing all over again

Date: Sun, 01 Jul 2018 15:36:23 +0000

To: notmuch@notmuchmail.org

Cc:

From: Leonard Lausen


I also experience this error trying to import the emacs-orgmode mailing
list archive. Has there been any progress on fixing the issue? What
steps should be taken to help?

Also it seems that the midsize corpus
https://giku.de/reduced-sample-midsized.xapian-exeption.tar.xz is not
available anymore.

David introduced a patch to improve the error message, below I paste the
message I get:

add_file: A Xapian exception occurred
A Xapian exception occurred adding message: Unexpected end of posting list for 'G00000000000255c6'.
Processed 4949 total files in 10m 13s (8 files/sec.).
Added 4921 new messages to the database.
Note: A fatal error was encountered: A Xapian exception occurred


Gregor Zattler <telegraph@gmx.net> writes:

> Hi David,
> * David Bremner <david@tethera.net> [14. Jul. 2017]:
>> Gregor Zattler <telegraph@gmx.net> writes:
>>> I redid the test on this mid-sized corpus and I got an
>>> exeption (again):
>> [snip]
>>> Error: A Xapian exception occurred. Halting processing.
>>> Processed 27611 total files in 21m 2s (21 files/sec.).
>>> Added 27481 new messages to the database.
>>> Note: A fatal error was encountered: A Xapian exception occurred
>>> [Inferior 1 (process 6752) exited with code 01]
>>> (gdb) 
>> 
>> I can finally replicate the crash. I do crash in a different place
>> (after processing 25184 files), so it suggests a certain non-determinism
>> or dependence on the host configuration.
>
> I would assume the first difference is in the file system:
> Perhaps the files are accessed in a different order!?
>
>> Anyway, now that I have a test case, I'll try and figure out what the
>> underlying Xapian exeception is.
>
>
> Thanks.
>
> Ciao, Gregor
> -- 
>  -... --- .-. . -.. ..--.. ...-.-
_______________________________________________
notmuch mailing list
notmuch@notmuchmail.org
https://notmuchmail.org/mailman/listinfo/notmuch

Thread: