Thank you for the heads up on the state of htMerge... We may be able to
help (here at ViWire) correct this if we could get more information about
the exact nature of the problem holding up the fix of htMerge.
It seems to me (based the little information I have) that mifluz is the
root of the problem. Could you or anyone who is involved with this part of
the code send me details about what mifluz does and/or what it is intended
to do? Is is a functional part of the code, or can it be safely extracted
from the mainline program? Is it easier to remove or to finish
integrating?
Thanks, and I'm looking forward to contributing to the quality of this
project!
-Rav
Gilles Detillieux
<[EMAIL PROTECTED] To: [EMAIL PROTECTED]
nitoba.ca> cc:
[EMAIL PROTECTED]
Subject: Re: [htdig] htMerge
Terminating Prematurely
10/17/2001 01:48
PM
According to [EMAIL PROTECTED]:
> I have been unable to get around this problem, and I can't find any
> documentation or explanation of this issue.
>
> For reference, the box I am running this on is a dual 550 Xeon, 10k rpm
> SCSI discs, 512MB RAM and 512MB swap.
>
> When I attempt to merge two indexes of 11,399 documents created with
htDig
> v 3.2.0b4.093001 (the command is: htmerge -v -c /etc/htdig1.conf -m
> /etc/htdig2.conf), the merge process begins running and almost
immediately
> I see a steady increase in the RAM consumption - until it consumes all
> available physical memory - then a steady increase in Swap memory
> consumption - until all available swap memory is also consumed. Then,
> after all memory is used up, htmerge rather suddenly stops and dumps me
> back to the prompt with a message - the single word "Terminated". No
> segmentation fault, no errors flashing by or even a status summary. Just
> the word "Terminated". After this happens, the database I was merging to
> appeards uncorrupted and still completely usable. However, when running
> htStat after the merge culminating in the "Terminated" message, I see
that
> the merged index has 17,247 documents (instead of the 22,798 it should
> have), which is suspicious since I know there are no duplicates between
the
> indexes.
>
> Any help in getting htMerge to work would be super-appreciated!
It's a known problem that htmerge is a memory hog in the 3.2 betas, but
unfortunately there's no solution yet. This code needs to be rewritten,
but if I recall correctly, Geoff had earlier explained that this rewrite
depends on first merging in the latest mifluz code, which has been stalled
for about a year now because the folks who first integrated mifluz into
htdig 3.2 aren't around to help out anymore.
--
Gilles R. Detillieux E-mail: <[EMAIL PROTECTED]>
Spinal Cord Research Centre WWW:
http://www.scrc.umanitoba.ca/~grdetil
Dept. Physiology, U. of Manitoba Phone: (204)789-3766
Winnipeg, MB R3E 3J7 (Canada) Fax: (204)789-3930
_______________________________________________
htdig-general mailing list <[EMAIL PROTECTED]>
To unsubscribe, send a message to <[EMAIL PROTECTED]> with a
subject of unsubscribe
FAQ: http://htdig.sourceforge.net/FAQ.html
