Menu

#270 htmerge 3.1.6, words.db=4GB corrupted

open-wont-fix
nobody
other (29)
5
2006-01-18
2006-01-18
No

It looks like words.db file can't be bigger than 4GB.

i use htdig 3.1.6 to index lots of documents (100 GO).

htmerge stops when words.db file is exactly 4 GB.
(other htdig files are > 4GO).

Is it a limitation of berkeley db 2.6.4?

Does htdig 3.2 have this limitation?

Regards,

Discussion

  • Gilles Detillieux

    Logged In: YES
    user_id=149687

    Yes, I believe it is a limitation of Berkeley DB 2.6.4, at
    least on some systems. We've had spotty reports of some
    people having success with database files larger than 2 GB
    on Solaris systems with ht://Dig 3.1.6, but database files
    larger than 2 GB are known not to work on Linux with this
    version. (You don't specify the platform on which you're
    running it.) Generally though, we don't recommend using
    3.1.6 with large database files, as it wasn't designed with
    those in mind.

    I'm pretty sure that ht://Dig 3.2.0b6 works fine with large
    database files on Linux as well as other systems.

     
  • Gilles Detillieux

    • status: open --> open-wont-fix
     
  • Gaetan QUENTIN

    Gaetan QUENTIN - 2006-01-19

    Logged In: YES
    user_id=799288

    It is on a Linux Mandrake 2006, kernel 2.6.12-12mdk .

    .docdb file is 1,1 GO
    .wordlist file is 7GO
    .wordlist.new file (sorted by sort) is 6 GO
    .words.db file is exactly 4 GO when htmerge stops.

    since i have to switch to htdig 3.2, is there a way to
    convert 3.1.6 files into 3.2 files, or am i obliged to
    launch again the htdig process, which took several days to
    build the files??

    Some guys talk about mysql instead of db2: is it possible?
    i have not found the patches to modify htdig 3.1.6 neither
    3.2...

    Regards,

     

Log in to post a comment.