Hi Ralf,

> Yeah, that should be understood as: Between 7:00 and 9:00 h in the morning the
> items are not indexed immediatelly, as I normally start TOSS manually at 9:00
> o'clock when opening my office.

OK, that makes sense.  If there is no connection from IQM to the indexer, the
IQM will back off for 30 seconds before retrying, so there should only be a
short delay after TOSS starts before those files get indexed and as you've
noticed, it's pretty fast.


> Antony, found some issues not answered - or at least I did not read and
> understood it correctly from the Admin Guide and the TOSS Setup Guide:
>
> 1. Is it a "must" to configure all parameters and add the missing ones in the
> TOSS config files? Some topics are not in the config file and need to be added
> manually, if they are used.

Maybe, can you clarify what ones you mean.  I think the minimum required is
simply to make sure the <service> element is configured with your mail/lib
service Id and to clone an existing <service> to add a new service.  There are
sensible defaults for most things.


> 2. After starting TOSS and indexing library service, how to handle entries
> remaining in the database\iq\qm096001.
>
> On my server. there is one file (f39vmkjv.002), dated 16.09.09, which is not
> indexed, but still remains?

New entries have the 'n' prefix and if the indexing fails for some reason, it is
renamed with an f prefix.  These items will never be handled again, so they can
be removed safely.  However, it indicates some failure.  Typically you will see
some indication of failure in the indexer's server.log and also in the
request.log.  There is an entry in request.log for each and every file sent from
database/iq/*.

Some examples of failure might be that the database thinks an attachment is
compressed, but it fails decompression.  The to.log will normally tell the type
of error that occurred.

> 3. There are lots of dot-failed files remaining in the TOSS directory under
> ..\ServiceData\9cpccwxr\0000000b\iq\001002 and ...\attach that have extension
> dot-failed.
>
> As far as I have investigated, 001002 holds XML-Files describing what to
index.
> The context to index is under \attach.
>
> If there are problems the extension will change by concatenating a '.failed'
to
> everything.

Correct.


> Can Administrators remove these files as a housekeeping to tidy up the
> directories?

Yes, these files can be safely deleted.

> Is it possible to start indexing of this files again to give it another try?

Technically yes, but you would have to shutdown toss and rename them to remove
the '.failed' and then restart toss.  However, unless there's a simple reason
for failure, it doesn't serve much purpose.  If you can identify the mail (from
the ID in the XML) you can just as easily rerun tomindex with the given ID,
which is probably as fast and less hassle!


> In general, the administration part telling "how to keep the TOSS up'n
running"
> in the Teamware Office environment is not described that much, is it?

No it is not - and we will look to improve that over time as we get more
customer feedback about typical things that happen in real life.

Many thanks for the comments
Antony
>
>
>
> Ralf Wieler wrote:
>> Hi All,
>>
>> two days ago, I installed a fresh TOSS service to my 7.4 Ed.1 Teamware Office
>> running on Windows Server 2003.
>>
>> Actually, it needs to be started manually.
>>
>> After taking indexing into use by modifying the config files to index Mail
and
>> Library Service, I seems that Library Service starts automatically to index
>> documents in the Libraries marked for text search using Windows Library
Client
>> --> Library Properties.
>>
>> But, within Mail Service, it only indexes new mails which arrives during the
>> time the TOSS was running after manual start.
>>
>> For my system, which is shutdown over night from 23:00 to 07:00 the result is
>> as follows:
>>
>> Let say, I start office work at 9:00 and first action after taking coffee is
> to
>> start the TOSS manually and keep it running. As a result only mails coming in
>> between 9:00 and 23:00 h will be indexed!
>>
>> Older mails, outgoing mails and mails coming in between 7:00 and 9:00 are not
>> indexed and therefore could not retrieved thus giving a poor search result.
>>
>> How to start the indexing/re-indexing of the Mail Service? Unfortunatelly,
>> documentation does not say much.
>>
>> Thanks for any advice and help
>> br
>> /Ralf
>>
>> Mit freundlichen Grüßen
>> Ralf Wieler
>>
>> ---  Your comments are very welcome  ---
>>
>> RW Consult IT-Dienstleistungen
>> Dipl.-Ing. Ralf Wieler
>> Zweitorstrasse 23
>> 41748 Viersen
>> Germany
>>
>> Tel.:   +49 2162 353580
>> mobil:  +49 172 2520437
>> Fax.:   +49 2162 353582
>>
>> Mail:     ralf.wieler@...
>> Internet: http://www.rwconsult.de
>
>
>
>
> ------------------------------------
>
> ---------------------------------------------------------
> TWOffice is an INDEPENDENT users' mailing list.
> More details at: www.proteusworld.com/twusersgroup
> Admin: Sebastian Paul Avarvarei
> Yahoo! Groups Links
>
>
>
>
>
>
> Mit freundlichen Grüßen
> Ralf Wieler
>
> ---  Your comments are very welcome  ---
>
> RW Consult IT-Dienstleistungen
> Dipl.-Ing. Ralf Wieler
> Zweitorstrasse 23
> 41748 Viersen
> Germany
>
> Tel.:   +49 2162 353580
> mobil:  +49 172 2520437
> Fax.:   +49 2162 353582
>
> Mail:     ralf.wieler@...
> Internet: http://www.rwconsult.de
> --------------
>
>
>
>
> ------------------------------------
>
> ---------------------------------------------------------
> TWOffice is an INDEPENDENT users' mailing list.
> More details at: www.proteusworld.com/twusersgroup
> Admin: Sebastian Paul Avarvarei
> Yahoo! Groups Links
>
>
>
>