big problem with loop-aes and reiserfs partitions

big problem with loop-aes and reiserfs partitions

[Newsmail]

hello everybody,
I'm using loop-aes in conjunction with lvm 1.0.6 and reiserfs. each 
partition has its own loop device, that is lvm-ed after, and on top of all 
I have reiserfs. I'm using 2.4.20aa1.
when I was using 2.4.19 I had no problem at all. when Oleg introduced the 
new block allocation algorithm in 2.4.20-preX I tried it out and I got the 
same problems that I'm having now with 2.4.20 final. In the 2.4.20-preX 
times I tried out this on multiple machines and all of them seemed to give 
the same results. so the problem is: when I'm copying huge amount of data 
on the encrypted filesystem first ssh seems to hang, I reconnect and the 
process that made the copying hung too. after all this, I am unable to 
delete the file that was being written. or at least I think so, I'm not 
sure. in conjunction with all this, the load is around 5. if I try this out 
again, it hangs again, and the load goes up to 10. and so on. and after all 
of this clear reboot is impossible, I have to make a cold reboot. I'm sure 
that drives doesnt have a problem at all, hardware is in good condition. In 
the past I was able to reproduce those problems on more than one machines, 
I'm sure I could do this this time, or that you will be able too :) or I 
hope so.
Please give me some directions how to solve this problem, or tell me what 
informations do you require more about the system I use, or what tests 
should I do.
sincerely,
greg

Re: big problem with loop-aes and reiserfs partitions

[Oleg Drokin]

Hello!

On Sun, Dec 22, 2002 at 10:21:37PM +0100, Newsmail wrote:
> when I was using 2.4.19 I had no problem at all. when Oleg introduced the 
> new block allocation algorithm in 2.4.20-preX I tried it out and I got the 
> same problems that I'm having now with 2.4.20 final. In the 2.4.20-preX 

Too bad you have not reported it back then.

> the same results. so the problem is: when I'm copying huge amount of data 
> on the encrypted filesystem first ssh seems to hang, I reconnect and the 

How much is this "huge"? Can it be compared in size with total fs capacity?
Is it one file of that size or is it lots of files of smaller size?

> of this clear reboot is impossible, I have to make a cold reboot. I'm sure 
> that drives doesnt have a problem at all, hardware is in good condition. In 
> the past I was able to reproduce those problems on more than one machines, 
> I'm sure I could do this this time, or that you will be able too :) or I 
> hope so.

I am going to try and reproduce it now. If I won;t be able to, I will
ask more questions.

Bye,
    Oleg

Re: big problem with loop-aes and reiserfs partitions

[Newsmail]

Yep, sorry I should have reported it back at that time :~
Well the huge amount isnt really 'huge' compared to the partition size. I 
copied 1 gig of data, composed of + - big files(10-60megs each) on/to a 
partition of size like 300gigs.
regards,
greg
At 08:18 2002. 12. 23.istá', you wrote:
>Hello!
>
>On Sun, Dec 22, 2002 at 10:21:37PM +0100, Newsmail wrote:
> > when I was using 2.4.19 I had no problem at all. when Oleg introduced the
> > new block allocation algorithm in 2.4.20-preX I tried it out and I got the
> > same problems that I'm having now with 2.4.20 final. In the 2.4.20-preX
>
>Too bad you have not reported it back then.
>
> > the same results. so the problem is: when I'm copying huge amount of data
> > on the encrypted filesystem first ssh seems to hang, I reconnect and the
>
>How much is this "huge"? Can it be compared in size with total fs capacity?
>Is it one file of that size or is it lots of files of smaller size?
>
> > of this clear reboot is impossible, I have to make a cold reboot. I'm sure
> > that drives doesnt have a problem at all, hardware is in good 
> condition. In
> > the past I was able to reproduce those problems on more than one machines,
> > I'm sure I could do this this time, or that you will be able too :) or I
> > hope so.
>
>I am going to try and reproduce it now. If I won;t be able to, I will
>ask more questions.
>
>Bye,
>     Oleg

Re: big problem with loop-aes and reiserfs partitions

[Oleg Drokin]

Hello!

On Mon, Dec 23, 2002 at 01:47:36PM +0100, Newsmail wrote:
> Yep, sorry I should have reported it back at that time :~
> Well the huge amount isnt really 'huge' compared to the partition size. I 
> copied 1 gig of data, composed of + - big files(10-60megs each) on/to a 
> partition of size like 300gigs.
> regards,

Ok, another question is where did you got your crypto api patches from?
Ones I took from kerneli.org cannot be compiled.

Bye,
    Oleg

Re: big problem with loop-aes and reiserfs partitions

[Newsmail]

http://loop-aes.sourceforge.net
I was using the newest version, with aes-128 encryption.
regards,
greg
At 15:20 2002. 12. 23.­, you wrote:
>Hello!
>
>On Mon, Dec 23, 2002 at 01:47:36PM +0100, Newsmail wrote:
> > Yep, sorry I should have reported it back at that time :~
> > Well the huge amount isnt really 'huge' compared to the partition size. I
> > copied 1 gig of data, composed of + - big files(10-60megs each) on/to a
> > partition of size like 300gigs.
> > regards,
>
>Ok, another question is where did you got your crypto api patches from?
>Ones I took from kerneli.org cannot be compiled.
>
>Bye,
>     Oleg