WriteLog
[Top] [All Lists]

[WriteLog] Save is Slow with large logs

To: <writelog@contesting.com>
Subject: [WriteLog] Save is Slow with large logs
From: w5xd@writelog.com (W. Wright, W5XD)
Date: Wed Mar 5 20:13:32 2003
There is one important exception to the statement that the 
_bck.adi file has enough information to restore the .wl file
after a crash: the WAE contest with QTCs. When WL logs receipt
or transmission of QTCs, it goes back to old entries in the 
logbook and updates them, and those updates are NOT in the _bck.adi
file. 

So, in the WAE, doing a File Save occasionally is necessary
as the QTC information is not on disk anywhere between File Saves.

Wayne

> -----Original Message-----
> From: writelog-bounces@CONTESTING.COM
> [mailto:writelog-bounces@CONTESTING.COM]On Behalf Of Eric Scace K3NA
> Sent: Wednesday, March 05, 2003 17:51
> To: Marty Tippin; writelog@CONTESTING.COM
> Subject: RE: [WriteLog] Save is Slow with large logs
> 
> 
> Hi Marty --
> 
>    Yes, I had the same problem on a 750 MHz machine in the WW 
> CW contest.
> 
>    After a series of email exchanges with Wayne W5XD about 
> the internal workings of WL and saving, it became apparent 
> that there was
> no need to "save" the log after each QSO.  If you want to 
> peel back the onion one layer, read on...
> 
>    After every QSO is logged, WL records it to disk in a file 
> named [name of your contest log]_bck.adi.  This disk file contains the
> additional QSOs made (anywhere in the network) since the log 
> was "saved" to the [name of your contest log].wl file.
> 
>    If the computer or WriteLog should crash, the operator 
> will be given the opportunity to pick up all the QSOs in the 
> _bck.adi file
> when he opens the log file.
> 
>    There is no material increased risk in losing the QSOs 
> made since the last save to the .wl file.
> 
>    In conclusion, a File/Save command simply integrates the 
> _bck.adi QSOs made recently with those already in the .wl file, and
> writes the now-larger .wl file back on the disk.  Then the 
> _bck.wl file is zero'd out, waiting for more new QSOs.
> 
>    This seems very secure -- even belt & suspenders 
> redundant, especially on a network with multiple computers, 
> each writing the
> full .wl and _bck.adi files on their respective disks.
> 
>    As a result, I've configure my WL to not save at all 
> (until I tell it to)... and not to give me a warning message 
> until I have
> arbitrarily-chosen 1000 new QSOs in the _bck.adi file.  
> During a slow moment in the contest I might do a File/Save 
> command, but I
> really don't worry about it anymore.
> 
>    Lastly, on a new computer with a 2.1 GHz processor and a 
> huge, empty hard drive, a 4500 QSO log saved in about a tenth of a
> second... so this confirms your observation that File/Save 
> time appears to be somewhat related to processor speed 
> (although faster
> disks are probably also correllated with fast processors).
> 
> -- Eric K3NA
> 
> -----Original Message-----
> From: writelog-bounces@contesting.com
> [mailto:writelog-bounces@contesting.com]On Behalf Of Marty Tippin
> Sent: 2003 March 5 Wednesday 17:31
> To: writelog@contesting.com
> Subject: [WriteLog] Save is Slow with large logs
> 
> 
> We used Writelog 10.39C last weekend in the ARRL DX SSB 
> contest Multi-Two
> effort from PJ2T. Computers ranged from a 450MHz Pentium II 
> to a 450 MHz
> Pentium III. Overall, the program worked great as it usually does.
> 
> However, there was a nagging problem which grew more 
> obnoxious with every
> QSO: After we reached somewhere around 2000 Qs, saving the log started
> taking longer and longer, up to about 10 seconds at the end 
> of the contest
> when we had 12500 contacts in the log. During the time the 
> log was being
> saved, the computer was worthless -- any Auto CQ would hang 
> the radio in
> transmit, and you couldn't type anything in the call entry window.
> Basically, we were forced to abandon the use of the DVK when 
> we thought an
> autosave might be coming up, and had to "fill" airtime while 
> waiting for
> the machine to become usable again.
> 
> The problem was most evident on the Pentium II machine, and 
> generally not
> an issue at all on the PIII box.
> 
> Anyone had similar experiences? Solutions? Suggestions?
> 
> Thanks
> 
> -Marty NW0L
>   martyt@pobox.com
> 
> _______________________________________________
> WriteLog mailing list
> WriteLog@contesting.com
> http://lists.contesting.com/mailman/listinfo/writelog
> 
> 
> _______________________________________________
> WriteLog mailing list
> WriteLog@contesting.com
> http://lists.contesting.com/mailman/listinfo/writelog
<Prev in Thread] Current Thread [Next in Thread>