backwards compat broken with R4-1.0.4

[Gentree]

I have a number of R4 partitions and currently use a 2.6.14 kernel .

I have reiser4progs-1.0.4 since I dont want to render the whole system so bleeding edge it cant be used with anything else.

However I now find that I cant even mount my new R4 partitions with a 2.6.11 kernel !!

I always like to have at least one fallback kernel incase I have issues to check or I have a breakage.

There is another major reason for keeping backwards compat of 1.0.4 : there are a number of system rescsue CDs that support R4 . They are currently useless to me.


Anyone looked into this ?

TIA, Gentree.

BTW I am aware that this kernel has 1.0.5 but that does not stop it mounting older partitions.
_________________
Linux, because I'd rather own a free OS than steal one that's not worth paying for.

KX7-333 , AthlonXP1800+ @2.3GHz
AthlonXP-M on A7N8X @ 2.6/2.4GHz (winter/summer)
2.6.32-hh1 : portage ~x86

[Rainmaker]

I noticed the same thing.

All kernels < 2.6.12 cannot mount a partition which has been mounted at least once with kernel >=2.6.12.

Only workaround I found this far (until livecd's support the newer format) is knoppix. Boot knoppix. Then use

apt-get dist-update
apt-get dist-update
apt-get reiser4tools

You should be able to fsck the partitions. GHranted, you still can't mount them, but at least you can repair them.
_________________
If you can't dazzle them with brilliance, baffle them with bullshit.

[Gentree]

Repair to what condition?!

That sounds like a recipe for disaster. Different kernel, different reiser4 modules arbitary update, different reiser4progs.


AFAIKR , there were serious issues on some platforms with 1.0.1 R4; there was no upgrade path from 1.0.3 to 1.0.4 ie format anew and copy across. Now 1.0.4 to 1.0.5 seems to be a one way street.


No, I think a large degree of caution is necessany here if you dont want to lose the whole partition.

I find this kind of forced migration very annoying. I really liked my 2.6.11-nitro2 kernel and I dont like only having one working kernel on a box.

I did not realise simply mounting a partition with an 1.0.5 capable kernel could alter the structure, Are you sure about that?

_________________
Linux, because I'd rather own a free OS than steal one that's not worth paying for.

KX7-333 , AthlonXP1800+ @2.3GHz
AthlonXP-M on A7N8X @ 2.6/2.4GHz (winter/summer)
2.6.32-hh1 : portage ~x86

[infirit]

[Gentree]

Thanks a lot for that link , my confusion is now well-founded and well-informed confusion

This flagrant lack of respect for any need of compatability between even minor build versions is lunacy. Once I have a clearer idea what is going on I will post to reiser ML about this.

This is exactly the sort of thing that has everyone flaming about R4.

The underlying fs is impressive despite it's youth but this sort of thing causes too much time wasting even for experienced users.

Since simply mount the device changes it , which would be completely unexpected to anyone who did not know I think this should have a big health warnig at the top of the 2.6.14-nitro2 thread.

Forwarned , is fore-armed.

It would make much more sense to know how to avoid and/or fix this problem before creating it by testing a newer kernel.

Ironically I had presensed something like this in seeing R4-1.0.5 in nitro and had put off installing this kernel for that precise reason.

So where do we go from here?

1/ abandon more recent kernels - is there a feature I need that justifies the breakage?

2/ accept I cannot use most liveCDs and even my very stable 2.6.11-nitro kernel and access from my dual-boot to suse9.1

3/ stop using R4

4/ try to find some way to make them cohabitate. - unlikely it seems.


Thanks again for you input.


_________________
Linux, because I'd rather own a free OS than steal one that's not worth paying for.

KX7-333 , AthlonXP1800+ @2.3GHz
AthlonXP-M on A7N8X @ 2.6/2.4GHz (winter/summer)
2.6.32-hh1 : portage ~x86

[Gentree]

OK, my choice was : 1.

I used the pre "pset" patched version of reiser4progs , ie the one I had before which has since disappeared from portage.

for reference it can still be grabbed from Gentoo cvs as version 1.5 of reiser4progs-1.0.4 ebuild.

I booted to my backup root partition on reiserfs using a 2.6.11 kernel. A few drives failed to load since not build against this kernel but I was able to login and run

fsck.reiser4 --build-fs

on all my R4 partitions. No errors other that what seemed to relate to the incompatiblilty after a bit of a wait everything came clean.

Now running 2.6.11-nitro2 with all my R4 apparently healthy .

So the underlying code seems vindicated and robust enough to withstand being mucked about with. That's very reassuring.

It would be even nicer if Herr Reiser would spare us these hair-raiser incompatibilities.
_________________
Linux, because I'd rather own a free OS than steal one that's not worth paying for.

KX7-333 , AthlonXP1800+ @2.3GHz
AthlonXP-M on A7N8X @ 2.6/2.4GHz (winter/summer)
2.6.32-hh1 : portage ~x86

[juliebread]

[infirit]

juliebread, Cry if you do not have any backups. However it is christmas in Russia so zam, vs are not able to check e-mails. Maybe they will have some good advice when they come back... monday I think.
_________________
EASY TO INSTALL = Difficult to install, but instruction manual has pictures.
Join the adopt an unanswered post initiative today

[Gentree]

First of all STOP and read up on the problem before charging ahead and maybe making a recoverable problem into a broken mess.

try this ... try this ... try this .....mentality will almost certainly leave you with nothing.

The problem is not so much reiser4progs - the userland utilities , but the reiser4 support you have in your kernel.

Re-read the whole thread carefully and evaluate where you are with respect to your kerenel and the utils.


_________________
Linux, because I'd rather own a free OS than steal one that's not worth paying for.

KX7-333 , AthlonXP1800+ @2.3GHz
AthlonXP-M on A7N8X @ 2.6/2.4GHz (winter/summer)
2.6.32-hh1 : portage ~x86

[juliebread]

I'm probably further confusing matters in this thread since I'm not running Gentoo. I did try posting to Debian's ML and Reiser's ML first, but received little help. My searches led me to this forum.

[Gentree]

[Gentree]

just out of curiosity what does the partition table look like from parted or fdisk?
_________________
Linux, because I'd rather own a free OS than steal one that's not worth paying for.

KX7-333 , AthlonXP1800+ @2.3GHz
AthlonXP-M on A7N8X @ 2.6/2.4GHz (winter/summer)
2.6.32-hh1 : portage ~x86

[juliebread]

You're right. This thread is maybe not dealing with the problem I have. I guess I was hopeful when you lead me from this thread over here.
http://forums.gentoo.org/viewtopic-t-366500-highlight-reiser4.html

But it doesn't appear the individual (Saundersx) who received the same error as I had his issue addressed.

[juliebread]

[Gentree]

Your're right , I think the key diff is that Saundersx had a backup

Since your prob is not due to the version changes dicsussed here, I suggest you stick with the same kernel version and reiser4progs so at least the two are in sync.

I had a similar disk full on my root partition recently due to misplacing where I was downloading a large ISO.

All went badly wrong with no space on root and I just powered off because it was getting very messy.

A moments reflection enabled me to realise the problem, I booted from a live CD, cleaned up and rebooted. No damage done.

I cant recall off hand if root was reiserfs or R4 at that point. I did post about it if you want to dig it out.



Dont worry about parted not knowing the fs , that's normal, its not R4 aware. If I see a blank with parted it usaully means it is R4 . Dont take too much notice of the 2MB used either, it's probably right in view of what fsck shows but dont rely on at partition tool for that sort if info.

[dreamer]

[juliebread]