Bug #230

Indexer drops information probably when creating summary

Added by Mark Schlieker 996 days ago. Updated 988 days ago.

Status:New Start:08/26/2009
Priority:Normal Due date:
Assigned to:- % Done:

0%
Category:Indexer
Target version:-

Description

I recently indexed a very large svn repository. It has over 50 GB of contents and several 100,000 files. It took about a week for the indexing process to finish and I noticed that the total size of the files in the index folder decreased dramatically in the end which makes me assume that a huge portion of information must have been lost.
I used "Luke Toolbox" to have a look into the index and noticed that several files which are present in repository can't be found in the index.
I let the indexer run scheduled by invoking supose in the following way (no further options provided):
supose schedule --configuration ./cfg_schedule.ini --configbase F:\svn_index

I used version 0.6.1-385

I also have a huge log file which I cannot provide directly because it contains sensitive information but if you like I can search for certain things to help reproducing the bug.

History

Updated by Mark Schlieker 996 days ago

I now rerun the indexing but without the scheduler option. I have the feelin it has to do with the task which creates the summary. I had over 10 GB of index after the index run approx. for 4 days and in the end it decreased to only 200 MB.

Updated by Karl Heinz Marbaise 995 days ago

Hi Mark,

can you post you ini file here ?

Kind regards
Karl Heinz Marbaise

Updated by Mark Schlieker 995 days ago

The ini-file is not really exciting. It looks like this: url = https://svn.domain.org/repos
indexusername = myusername
indexpassword = mypassword
fromrev = 0
torev = HEAD
resultindex = summary
cron = 0 30 13,22 * * ?

/>

Updated by Mark Schlieker 995 days ago

Mark Schlieker wrote:

The ini-file is not really exciting. It looks like this:

The format got lost, here it is preformatted:

[PROJECTXY]
url = https://svn.domain.org/repos
indexusername = myusername
indexpassword = mypassword
fromrev = 0
torev = HEAD
resultindex = summary
cron = 0 30 13,22 * * ?

Updated by Mark Schlieker 989 days ago

Hi.

When running the indexer without scheduler the index is created properly.
There is no summary then but I think the summery is only created for the scheduler, isnt it?

Regards,
Mark

Updated by Karl Heinz Marbaise 988 days ago

Hi Mark,

When running the indexer without scheduler the index is created properly. There is no summary then but I think the summery is only created for the scheduler, isnt it?

The idea with the scheduler is to create a new index with the new revisions (e.g. every hour or so) and to merge the result index with the created one...
So the final result should be to have a summary where the old information is stored of the previous run and the new one...
But it seemed to me that there is a bug in ...

Thanks for reporting this...
I have to take a deep look into it...

Kind regards
Karl Heinz

Also available in: Atom PDF