The Linux 4.0 Kernel Currently Has An EXT4 Corruption Issue

[Isedonde]

I have to agree, it's probably a symptom of less-than-stellar kernel testing procedures. Everyone just sends patches, hopefully tested well enough, then the guys on higher levels of the commit hierarchy take a look at the code, and finally Linus says "just a usual release, some changes here some changes there, please test" and then it's released eventually.

So no formal testing, just random people who may or may not use different features of the development kernel and if there are no complaints after some time, it goes public as a "stable" release. Not very convincing.

Some time ago I was unsure if I should really use BTRFS raid1 for my file server. I almost decided to use ext4 raid1 instead. Is ext4 on a raid1 volume affected by this bug? If so, I'm glad I use BTRFS ? the checksum feature appeared to be worth it.

[Mike Frett]

NTFS is rock solid and has been for more than a decade.

Anything that requires me to defragment it at least once a month in order to have stable performance; isn't something I could call rock solid.

[R00KIE]

I think there is something in common in the systems affected by this, at least two of them use SSDs. There is currently a growing blacklist of SSDs that corrupt data when queued trim is used so I suppose the problem could be caused by that.

There is also this: http://www.gossamer-threads.com/list...readed#2176274

[gamerk2]

Anything that requires me to defragment it at least once a month in order to have stable performance; isn't something I could call rock solid.

Funny, I haven't defragged a NTFS drive in years. Granted, that's largely because HDD space isn't really an issue anymore, and with 8GB RAM kinda standard now, it's not like HDD activity is going to be a bottleneck.

Seriously though, for a stable release, this is unacceptable. I work for an aerospace corporation as a SW Engineer, and about 10% of what I do is coding. The other 90% is testing and documentation of said testing. HDD corruption in stable builds went out of style 20 years ago. Stop making excuses and fix your testing methods, because whatever you guys are doing clearly isn't working.

[gens]

Only an idiot would run a very little tested kernel in a production enviroment.

i think you underestimate the number of idiots working in IT
to be fair, they are too lazy to upgrade to latest

[pal666]

It's fun to see the Linux zealots in this thread trying to defend the indefensible. Linux 4.0.2 is considered a stable kernel, so stop saying things like "you should not use it in production" as a lame excuse for the horrendous development model. NTFS is rock solid and has been for more than a decade.

it is fun to see windows zealots coming to linux forums and pulling fud out of their asses. linux 4.0.2 is not shipped in rhel, which is an example of stable distro, so you could continue to suck dick silently. btw, more than decade old ext4 is called ext2, does it have corruption in 4.0.2? did you pay linus money for giving you tested kernel ? no. you pay google for android or redhat for rhel. their kernels are tested. 4.0 is tested on gentoo users. dont like it ? don't be gentoo user.

[pal666]

Seriously though, for a stable release, this is unacceptable. I work for an aerospace corporation as a SW Engineer, and about 10% of what I do is coding. The other 90% is testing and documentation of said testing. HDD corruption in stable builds went out of style 20 years ago. Stop making excuses and fix your testing methods, because whatever you guys are doing clearly isn't working.

then stop making bullshit claims. if you work for aerospace then you don't download kernel from linus, you get kernel from rhel which is 3.10 for rhel 7. see, 3.10 is ten releases behind 4.0 ? did you pay linus money for giving you tested kernel ? no. you pay google for android or redhat for rhel. their kernels are tested. 4.0 is tested on gentoo users. dont like it ? don't be gentoo user.

[pal666]

http://linuxfonts.narod.ru/why.linux.is.not.ready.for.the.desktop.current.htm l

according to url you are some russian idiot

[ngkaho1234]

It's fun to see the Linux zealots in this thread trying to defend the indefensible. Linux 4.0.2 is considered a stable kernel, so stop saying things like "you should not use it in production" as a lame excuse for the horrendous development model. NTFS is rock solid and has been for more than a decade.

hilarious! lol

[Nille]

Anything that requires me to defragment it at least once a month in order to have stable performance; isn't something I could call rock solid.

Every Hard Disk fragment over the time. Ext specialy if you are on low disk space or if the volume is old and you write and delete many files.

Funny, I haven't defragged a NTFS drive in years.

But the planed Task in Windows does it. Since Vista there is a Task that Check and Defrag the Disks each Week. That happens also if the PC is not used and you see it that after some minutes of idle that the disk activity increases until you do something.