Hey guys,
Anyone tried etx4 yet? I'm downloading Ubuntu 9.04 right now... I'll tell you how it goes...
I've been using ext4 on a Fedora 11 installation in VMWare for a while now. No problems so far - it just works.
Some people had issues with GRUB and ext4 so they used ext3 for /boot. I haven't been able to test it myself.
ext4 had some issues during development including a filesystem corruption bug, however the bugs have been resolved and it should be ok to use.
Regardless of your filesystem, you should always backup your stuff
I dunno... but it seems a tad faster...
I know.
My VM machine of 9.04 is faster than my Vista native installation.
Can't se any reason why it wouldn't be, except disk performance will be lower than on a native installation (it's still on an NTFS partition). It all depends what hardware you "gave" it.
Yeh, I dunno... can't remember what I gave my Ext 3 machine.
No, I'm keeping EXT3 for the rest of this year.
ext4 is running fine on kubuntu, arch, and LFS at the moment (had to pull the LFS from the closet so to speak and install ext4)
though i still use ext2 for my /boot
| amdfangirl wrote : I know.
|
Doing what? How did you measure it??
I'm not 100% accurate, it's just how it feels.
Can you guys vacuum down here and turn some lights on please?
It's pretty dusty and there is a bad echo.
heh heh heh ....
/runs
back to the other, reynod!
Actually I have a Linux firewall here ... Smoothwall.
I had IPCOP before.
I just had to get my Penguin mate Craig in to do the fiddly bits to get it going.
Linux doesn't often break like WinDoze.
A well configured and properly managed Linux system can stay up without any software failures for many years.
It is not unusual to see Linux servers with uptimes of 5 or 6 years or even longer.
Of course the owners of those systems should probably be updating their kernels at least to apply security patches.
| Quote : Of course the owners of those systems should probably be updating their kernels at least to apply security patches. |
The kernel can be patched without rebooting, so patching isn't incompatible with long uptimes.
I'm assuming you're referring to kexec, if that's the case you must have kexec support, sane hardware and even then it may not work
YMWV
I was actually thinking of KSplice. Not perfect, but could save a lot of reboots. I think for a truly no downtime situation you have to run mirrored or clustered servers. That's what I did when we wanted 100% uptime on a Citrix installation.
I'm guilty of running a Windows Server...
Too bad it has to be up to constitute it as running.
So that's...
Uptime?
They've not restarted our main windows db server in 6 months, and when they did it was only due to our sw needing it to complete the install. The Sun box on the other hand has been restarted 6 times in the last two weeks. Go figure?
There have been various reports of Unix and Linux servers being forgotten, having concrete walls built around them during renovation and reconstruction projects and still staying up and operating properly for 10-12 years.
This demonstrates how stable and reliable the operating system can be as long as the hardware is stable.
They sure don't build hardware like they used to, that's for sure.
Much to the disgust of an old workmate and then house mate I once threw a Solbourne dual proc spark workstation from a second floor window into a skip during a clear out of the comms room. Whilst the case suffered he did get it to post when he dragged it out from under the two generic servers I threw on top of it later that day. Shame he never managed to get an OS on it (it was a STRANGE beast) but it did make a coffee table in the house for a while.
I'm sticking with ext3 until Fedora13
(lucky number)
I'm still waiting for ext4 to have live-defrag implemented.
Had some corruption with Ext4, but i believed this is fixed now. Still i think it was integrated too early in Ubuntu and 9.10 making it default sure isn't the safest choice i've seen from the debian keeping-you-in-the-stoneage mentality. So some raised eyebrows here.
But generally i like ext4 alot better than ext3. That said ext3 just sucks, XFS, JFS, ReiserFS, mostly any FS is better than Ext3. And ext3 has some serious issues with maintenance and reliability; i consider it comparable to FAT of earlier Windows 98 age. Sure its (light) journaling, but it has many weaknesses and relies on checking the volume for errors rather than being a self-healing filesystem.
Ext3 is decent. Better than NTFS, less fragging to do.
Wouldn't mind ZFS as an option tho.
| amdfangirl wrote : Ext3 is decent. Better than NTFS, less fragging to do.
|
Isn't the problem with ZFS on Linux that it is proprietory? Works a treat on FreeBSD.
Not sure.
I really can't remember.
yeah ZFS is not license compatible with linux so the only way to get it to work is through fuse which makes it not worth the time
though you could try BtrFS (unstable at the moment). It's suppose to have ZFS like functionality
Btrfs does look very interesting, but not ready for the prime time just yet.
very true, definitely not for prime time yet, though a ZFS like native file system would be very nice on linux
Ext4 isn't ready either.
Still developing really...
I had some problems with an ext4 partition on a reinstall: a fix had made it into an updated kernel and reverting to the original one made a mess of it.
Personally, I'll stay on ext3 until btrfs becomes stable. One 'experimental' FS in use at once, KTHXBY.
I'd VM unstable stuff.
Yeah, that too - but then in order to really test a FS you need to use it, and thus you'll backup your sensitive data anyway - might as well do so on a live, spare system then.
Still, that's a good idea.
Its all Oracle's fault.
Though maybe Neelie Kroes can save us, because she still has to authorize the acquisition of Sun Microsystems by Oracle. Oracle is a very .. destructive company, and has destroyed alot of potentially sound technology in the past.
I reckon!
The merger was a horrible idea.
There are 1351 identified and unidentified users. To see the list of identified users, Click here.

