Continuous loop

[jeroen]

With virus scanner enabled, JkDefrag 3.7 finished the drive on which the ISO was present within 10 minutes.

Then there is only 1 conclusion possible, your virus scanner has a bug. It somehow impedes disk access when it is supposedly turned off. Which virus scanner are you using? So I can tell other users with the same problem.

[jtulp]

Sorry??? Let me sum it up again:

JkDefragx64 3.8 uses 50% of my CPU without any significant disc activity while moving a file greater than 3GB, regardless if my Virus scanner is actively scanning or not (checked both). When I choose to stop JkDefrag it is still active after 2 hours of CPU-usage by the program. Killing the process(-structure0 does not work and also shutdown or restart of my pc can not be performed.

JkDerfragx64 3.7 finishes defragmenting my drive within 10 minutes with my Virus scanner active.

For what it is worth, I use Avast Home Free! 4.7.1001 and when I disable scanning, I select to disable all modules (not only on-access scanning). I have a Core2Duo E6800, 2GB PC6400 and a SATAII 160GB/16MB Western Digital HD.

[jeroen]

JkDefragx64 3.8 uses 50% of my CPU

I don't want to quibble, but as I said before, it is not JkDefrag that is eating your CPU. It has handed a command to Windows and is not doing anything. The activity and CPU that you see is Windows executing the command, not JkDefrag.

JkDerfragx64 3.7 finishes defragmenting my drive within 10 minutes with my Virus scanner active.

I have taken a look at the sources, but I see no differences between v3.7 and v3.8 that could explain the behaviour you are experiencing. It's good that you have found a solution for your problem yourself (not turning off the virusscanner), because I have no idea what I can do in JkDefrag to fix your problem.

when I disable scanning, I select to disable all modules

Yes, but the virus scanner still has it's hooks into Windows. When you disable it then it's supposed to pass everything, but doesn't. It is still interfering with Windows somehow.

[jtulp]

Ok, I will promise I will make another test with 3.8 and first stop my Virus scanner services and killing any ash* process which is left and if it still occurs, then I am out of options. Off course my ISO's are now not fragmented anymore, so if it does not occur, then it could be because of this. Another test after reboot and maybe a copy back and forth of the ISO should give me more information on this.

[jtulp]

On one hand I was hoping you were right as this issue is getting on my nerves and I would like to have found the cause no matter if it was me at fault or not. On the other hand I was sure, that I was not mistaken and here againmmy results:

I stopped all modules for Avast, stopped the remaining 2 services and finally killed the systray icon (no other way to get it to quit normally). After this I started JkDefrag for my drive with the big ISO-files and after about 10 to 30 seconds of HD activity it locks up with the already mentioned symtomes. I created a log file again only this time with level 2, so not too much details shown extra (however the already posted log would most likely be the same).

Code:

00:00:00 JkDefrag v3.8, (c) 2006 J.C. Kessels
00:00:00 Date: 1900/01/00
2007812845:00:00 Starting defragmenter for 'C:'
2007812845:00:00 Opening volume '\\?\Volume{XYZ}' at mountpoint 'C:'
00:00:00 Step 1: Analyze
00:00:00 FreeSpaceTop=234563
00:00:00 Step 2: Free space
00:00:00 Step 3: MFT zone
00:00:00 Step 4: Directories
00:00:00 Step 5: Defragment

If I check the known changes in 3.8, then I can only imagine that the first mentioned bugfix (for infinite worm) might be the cause.
If you would be so kind to compile a JkDefragx64 for me to test with, without this change, then I would be more then happy to test if this resolved my issues (so we might have pinpointed the problem).

[jeroen]

Thank you very much for all your efforts in trying to help me fix this problem. It looks like it is not caused by code inside JkDefrag, so at the moment I am stumped. Maybe I'll get a bright idea in the future.

[jtulp]

Ok, well I am in doubt. Currently back on using 3.7 but I think I might look into something else in the near future.

JkDefrag is a great program in potential, but suffers from a few problems which are not aceptable to me. What I do not understand is, that if version X works and version X+1 does not work in an isolated environment (reproduction of a situation), that you blaim external factors. The issue can only be in 1 place and that is inside that specific program.

[jeroen]

What I do not understand is, that if version X works and version X+1 does not work in an isolated environment (reproduction of a situation), that you blaim external factors.

I can see your reasoning, but I don't see what else I can do but to blame "external factors". JkDefrag executes the following command at the point you are experiencing the problem. This command calls the Microsoft defragmentation API and instructs it to move a file from one place on disk to another. The command will return when it has finished with the move.

Code:

 ErrorCode = DeviceIoControl(Data->VolumeHandle,FSCTL_MOVE_FILE,&MoveParams,
    sizeof(MoveParams),NULL,0,&w,NULL);

The command works fine if you have activated your virus scanner, and is extremely slow when you deactivate it. To me that is positive proof that your virus scanner is buggy and is somehow interfering with Windows. The problem is NOT caused by JkDefrag code being executed (for example a loop), so there is nothing I can do. Please please please contact the makers of your virus scanner!

[jtulp]

I admit, that the virus scanner plays a part and causes some delay (as it does in every operation which reads from HD). I did no comparison in speed with virus scanner on or off, I just stated, what my pc does: it has a bit of HD-activity and after 30 seconds it stop. Whereas you can expect a lot of HD-activity while an defragmentation tool is active.

With all due respect, I made some mistakes myself maybe in explaining the situation. Although I respect your work, as I stated JkDefrag is great beside the mentioned problem. You fail to recognise to the issue at hand. The Virus Scanner was brought up in your first reply, however my tests have shown, that the Virus Scanner is irrelevant. Issue does not occur under 3.7 and Virusscanner enabled/disabled has no impact on functionality of 3.8. The tests I did, were to isolate the problem and rule out all the unknowns as much as possible.

I like your program, but am currently restricted to using 3.7, satisfactory by the way. However your response to this issue has a bit clouded my judgement and reduced my trust in the program. I do not demand anything, I just expect you treat this issue with the care and thought as you have put into JkDefrag.

[jeroen]

I have scrutinized the changes between v3.7 and v3.8, but at this time I do not see what could be causing your problem - 100% cpu and no disk activity, and only a single "moving clusters from to" line in a level 6 debug log.

[jtulp]

The same issue also occurs with 3.7 now. Strangely I ran 3.8 several times (both before and after a run with 3.7, virusscanner enabled/disabled) and everytime it hang. Now I have run 3.7 approximately 10 times it locked up once in exactly the same way as 3.8. So it seams that I was too quick in my judgement that it did not occur with 3.7 how strange it even may be that it only appeared just now and not sooner.

[jeroen]

Looking again at the log you posted earlier I now notice that some lines have a weird timestamp, and the others have a zero timestamp. I assumed before that you had changed the times, you said you had changed some things, but let me ask you to make sure. Did you change the times? If not then there is something wrong with the way JkDefrag accesses the clock on your computer. It uses the clock in a number of ways, most notably to control it's speed, so this might explain why JkDefrag is behaving erratically on your system.

[jtulp]

I only changed, file disc and volume information for privacy reasons. All other information is directly from the logfile. I use Vista x64 on an Intel Core 2 Duo with an Asus P5B motherboard.

Legal Vista by the way. When I bought a new one pc, I bought Vista Home Basic (just to rule out the thought of those Time Stoppers for postponing the activation that I used prior to that).

[jeroen]

I've made a special version for X64 with a fix that I hope will solve your problem. Are you willing to test it for me and let me know if the date/time is now correct? You can download from:v3.8.1 for X64

If this does not fix the problem could you then run the program with the "-d 6" debug option? The logfile will be huge, but I'm only interested in the last part where it is moving the big file. Let the program run until completion, do not stop it.

[jtulp]

Off course I will test. Thank you, I will report back somewhere during the day.

Was able to check by VNC just now (8:30) and it still occurs:

Code:

07:03:09 JkDefrag v3.8.2
07:03:09 Date: 2007/05/15
07:03:09 Starting defragmenter for 'C:'
07:03:09 Opening volume '\\?\Volume{XYZ}' at mountpoint 'C:'
07:03:09 MftStartLcn=786432, MftZoneStart=3359040, MftZoneEnd=3410272, Mft2StartLcn=4194969, MftValidDataLength=13424
07:03:09 MftExcludes[0].Start=786432, MftExcludes[0].End=799856
07:03:09 MftExcludes[1].Start=3359040, MftExcludes[1].End=3410272
07:03:09 MftExcludes[2].Start=4194969, MftExcludes[2].End=4208393
07:03:09 Step 1: Analyze
[...]
07:05:56 Gap found: LCN=14552308, Size=199308
07:05:56 Gap found: LCN=14752547, Size=168853
07:05:56 Gap found: LCN=15013673, Size=8415111
07:05:56 D:\X\Y.iso
  Moving 1140849 clusters from 14921400 to 15013673.

(I disabled/killed my virusscanner completely)

There might also be a minor cosmetic issue: the last line in the log is also shown on the screen, however 15013673 is cut of at the right edge as 150136. I use some personal font-tweak for Vista (not that rigoreous as those posted on tweak-sites), but this might cause this issue. My reporting is just a heads-up, if anyone else reports it as well, then you might consider it an issue.