Slower performance with ext4

[Chan Chung Hang Christopher]

fyrbrds at netscape.net wrote:
>  Good references given. Google returns a lot of corruption hits but most are from Jan-March of this year. Those are related to write-caching which is always a risk in case of power failure or system hang (mentioned in most that I read). There was reference to a boot-time mount parameter as a "work around," which may have to do with disabling write-caching. This would definitely affect performance. I am still reading docs to see if that is indeed what has happened.
>
>
>
> The document https://bugs.launchpad.net/ubuntu/+source/linux/+bug/453579 makes for very good reading (if you can stay awake). It includes discussion of whether or not to use ext4 as default in this release or revert to ext3. Some people wanted to stick to the traditional Debian conservatism and others (the Mavericks) sided with the bleeding edge. This was enlightening. The case was made that it was better to not take any chances and just use ext3. Then it was pointed out that of the corruption bug reports, there was no definitive link between the reports and they could not reproduce the error consistently enough to say that this component or that component/driver was defective. I found this to be the most revealing message:
>
>
>
>
>
>   " The relation to that upstream bug is tenuous at best.  The upstream bug:
>
>    
> - is reported against a newer kernel than the one we're shipping
>
>    
> - is reported to only happen when ext4 is on top of the DM layer, whereas Scott's
>       case was ext4 on a raw device
>
>    
> - is reported in connection with an unclean shutdown and subsequent
> fsck,
>       whereas Scott reported corruption of files without an unclean
> shutdown 
>       (but no mention in this bug of whether the corruption requires
> an intervening 
>       reboot/fsck to appear - Scott, please clarify)
>
>
>
>
>    So that upstream bug link should be dropped; it really doesn't look like the same bug."
>
>
>
>
>
>
>
>  So this was the reason for going ahead with the ext4 deployment depending on how you interpret what is being said here. I have to say that given how Linux is used and considering these docs, sticking with ext4 was a risky move. The decision may be vindicated if no bug is found with the file system itself. Distros always have to be wary of being last to roll out a feature. Why? Two words: Competitive Disadvantage.  I can't imagine that this is an easy decision for the guy who has to balance those two forces. I have to think that EULA's which absolve software companies of everything but dodging taxes play a big role in how these decisions are made, nicht wahr?
>
>   
So long as we get to choose what filesystem is used, I really do not 
care what they decide is default. jfs
has the second best write performance since the last test I saw and it 
has proven itself recently by surviving a ntop, no filters applied, rrd 
database creation run involving thousands of hosts without crashing the 
system nor loosing data. I will just stick with what is proven and still 
performing well.