I have ran memtest86 over night without an error. (http://www.memtest86.com/)
With that out of the way, I am having a problem with Par2's that repair, but they say they need to be repair. Or they repair, but Rar files, and checking the with QuickPar again shows they are bad. Or repeatedly pressing F5 and Repair will finally repair them and they can be extracted by RAR, only to have QP say there is an error if i hit F5 after successfully extracting them.
Making a .SFV shows that most of the .X files that QP creates are the same, but not all (2-9 were the same, the .1 and .rar were different. And making a .SFV from 2 different sets of files that extracted showed that one of the 4 files had differing CRC values, yet RAR had no problem extracting them.
Does this sound familair to anyone? The whole problem sounds like memory, except i am not having other memory problems. All day I have been running eMule, NewsPro (2 memory hogs) and compiling with VB. My memory has been getting a workout.
In my instance I am saving my files to an external USB2 Hard drive and running QP on those files when i get a problem.
I am wondering if anyone is having similiar problems and is using a USB drive as well.
Regards.
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
Yes, have run into the similar problem several times. Few times I had to try and try and try about a dozen times for it to fix the problem. Once I gave up after almost 2 days trying to get it to rewrite the rar.1 files.
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
Though you may have good memory, PAR2 will fail if it was created on a system with bad memory.
If you have lots of extra time and recovery data you might try forcing recovery with the files of your choice. Assuming that just a few PAR2 files have been created incorrectly, you might try moving them away one by one until you find which one is bad. You might be able figure out more quickly which PAR2 files are bad if there is disagreement on which original files are damaged.
Perhaps there needs to be better checks on PAR2 file integrity.
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
I have a similar case here, a 90x50 archive that is missing 8 files (they are completely missing as on the web interface we are downloading from files are either their or not there), "covered" by 9 PAR2 files with a sufficient number of blocks. I tried on 3 different machines (with QuickPAR 0.5.1 and 0.7.5) and on all 3 machines the end result was the same: after running for more than 4 hours (on a 3 GHz P4!) all the "restored" files are bad. All 3 machines haven't shown the slightest problem that might indicate a memory error (and they are in use quite a lot and quite heavily for video encoding and dvd authoring).
I also want to add that for people downloading from web interfaces par2 is a very bad invention, par1 was good enough for that (I understand that PAR2 has advantages on NNTP servers but if whole files are completely missing then unreliable file restoring and much longer time for restoring are the only things PAR2 has brought for us, the smalelr the block size the worse it gets).
One more thing:if the person who created the par2 files had faulty memory then why does QuickPAR not detect that? Shouldn't it detect faulty par2 files the same way as it detects faulty blocks in the files from the archive?
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
> One more thing:if the person who created the par2 files had faulty
> memory then why does QuickPAR not detect that? Shouldn't it detect
> faulty par2 files the same way as it detects faulty blocks in the
> files from the archive?
par2 files contain two types information:
1) Verification data;
2) Recovery data.
A memory error during the creation of the par2 files could affect either of these.
With every version of QuickPar since it was first released (i.e. all versions up to 0.7.5), when the creator uses the Verify button, the Recovery data does not get tested. The only way they can be certain that the Recovery data in the par2 files is good is to delete some of the data files and run a repair.
Version 0.8 of QuickPar (which has not yet been released) includes extra code in the par2 creation routines that is designed to detect corrupt Recovery data before it is written to the par2 files.
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
I've downloaded a file made up of 51 Rars, but 8 of them are missing.
However, there are a number of files with the extension .par2.
The first file is vpc.par2, but the other par files have a vol extension too, ie vpc.vol000.01.par2, vpc.vol131.14.par2 etc
When I run the vpc.par2 file, it says in the dialogue box, that there are no recovery blocks, but if I run an individual par file with the added 'vol', it says I have a number of blocks correspondng to the last two digits of the file.
Any suggestions on how to link all the par2 files together ? Or am I just doing something mundanely stupid....
Any help would be most appreciated....
Thanx..... Delboy
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
The PAR2 files have been renamed from what they should be so they are not being automatically recognised. Just use the Add button to have QuickPar scan them.
If you would like to refer to this comment somewhere else in this project, copy and paste the following link:
I have ran memtest86 over night without an error. (http://www.memtest86.com/)
With that out of the way, I am having a problem with Par2's that repair, but they say they need to be repair. Or they repair, but Rar files, and checking the with QuickPar again shows they are bad. Or repeatedly pressing F5 and Repair will finally repair them and they can be extracted by RAR, only to have QP say there is an error if i hit F5 after successfully extracting them.
Making a .SFV shows that most of the .X files that QP creates are the same, but not all (2-9 were the same, the .1 and .rar were different. And making a .SFV from 2 different sets of files that extracted showed that one of the 4 files had differing CRC values, yet RAR had no problem extracting them.
Does this sound familair to anyone? The whole problem sounds like memory, except i am not having other memory problems. All day I have been running eMule, NewsPro (2 memory hogs) and compiling with VB. My memory has been getting a workout.
In my instance I am saving my files to an external USB2 Hard drive and running QP on those files when i get a problem.
I am wondering if anyone is having similiar problems and is using a USB drive as well.
Regards.
Have you tried checking/repairing them with the command line version of PAR2?
Yes, have run into the similar problem several times. Few times I had to try and try and try about a dozen times for it to fix the problem. Once I gave up after almost 2 days trying to get it to rewrite the rar.1 files.
Though you may have good memory, PAR2 will fail if it was created on a system with bad memory.
If you have lots of extra time and recovery data you might try forcing recovery with the files of your choice. Assuming that just a few PAR2 files have been created incorrectly, you might try moving them away one by one until you find which one is bad. You might be able figure out more quickly which PAR2 files are bad if there is disagreement on which original files are damaged.
Perhaps there needs to be better checks on PAR2 file integrity.
I have a similar case here, a 90x50 archive that is missing 8 files (they are completely missing as on the web interface we are downloading from files are either their or not there), "covered" by 9 PAR2 files with a sufficient number of blocks. I tried on 3 different machines (with QuickPAR 0.5.1 and 0.7.5) and on all 3 machines the end result was the same: after running for more than 4 hours (on a 3 GHz P4!) all the "restored" files are bad. All 3 machines haven't shown the slightest problem that might indicate a memory error (and they are in use quite a lot and quite heavily for video encoding and dvd authoring).
I also want to add that for people downloading from web interfaces par2 is a very bad invention, par1 was good enough for that (I understand that PAR2 has advantages on NNTP servers but if whole files are completely missing then unreliable file restoring and much longer time for restoring are the only things PAR2 has brought for us, the smalelr the block size the worse it gets).
One more thing:if the person who created the par2 files had faulty memory then why does QuickPAR not detect that? Shouldn't it detect faulty par2 files the same way as it detects faulty blocks in the files from the archive?
> One more thing:if the person who created the par2 files had faulty
> memory then why does QuickPAR not detect that? Shouldn't it detect
> faulty par2 files the same way as it detects faulty blocks in the
> files from the archive?
par2 files contain two types information:
1) Verification data;
2) Recovery data.
A memory error during the creation of the par2 files could affect either of these.
With every version of QuickPar since it was first released (i.e. all versions up to 0.7.5), when the creator uses the Verify button, the Recovery data does not get tested. The only way they can be certain that the Recovery data in the par2 files is good is to delete some of the data files and run a repair.
Version 0.8 of QuickPar (which has not yet been released) includes extra code in the par2 creation routines that is designed to detect corrupt Recovery data before it is written to the par2 files.
I've downloaded a file made up of 51 Rars, but 8 of them are missing.
However, there are a number of files with the extension .par2.
The first file is vpc.par2, but the other par files have a vol extension too, ie vpc.vol000.01.par2, vpc.vol131.14.par2 etc
When I run the vpc.par2 file, it says in the dialogue box, that there are no recovery blocks, but if I run an individual par file with the added 'vol', it says I have a number of blocks correspondng to the last two digits of the file.
Any suggestions on how to link all the par2 files together ? Or am I just doing something mundanely stupid....
Any help would be most appreciated....
Thanx..... Delboy
The PAR2 files have been renamed from what they should be so they are not being automatically recognised. Just use the Add button to have QuickPar scan them.