[ various snips ] On Thursday 25 December 2003 07:26, Spider wrote: > begin quote > On Thu, 25 Dec 2003 09:32:20 -0500 > > Robert Crawford <[EMAIL PROTECTED]> wrote: > > On Thursday 25 December 2003 8:43 am, Spider wrote: > > > Not really, overall I advice against ReiserFS because of their > > > horrid recovery-tools. > > > Ext3 isn't the fastest in the race, but it has a darn good support > > > team. > > > That matters a lot for me.
> > > Reiserfs has a -horrid- way of dealing with the processor, and requires > a lot of cpu munching. > > Reiser has its uses ... But I wouldn't put > it even near a partition with data I value. it may be "stable" for > users, and so on. that doesn't matter if the fsck tools are so horridly > handicapped that they cannot recover data without forcibly rebuilding > the tree. > > As said, I ran reiserfs for a while, but won't ever do it again because > of how it cannot recover when it goes down the drain. > > put short: Reiser has performance, but uses CPU power more than > anything (not good when compiling). And it can't recover data once > things really go down the drain. > Sort of sums up my most recenta experience. I didn't even suffer a power outage. I deleted an reorganized some partitions following the resier partition (root partition for a SUSE 9.0 system), then shutdown normally. When I booted again, reiser believed that the partition origin had changed (it had not; same starting/ending cylinder as before), refused to decode the super block, marked the partition as readonly, and subsequently failed the rebuild tree, etc. Nothing would make th journal usable again. Fortunately most of the daa in my /home directory was intact, so I was able to copy off the data I needed and to reinstall using ext3 without a major loss. Needless to say, this experience (never encountered using ext3) left a sour taste in my mouth, It may well be that I don't understand reiser well enough and that I made some simple screwup that contributed to the problem, but I, too, would "never again put it near data I value." -- Collins -- [EMAIL PROTECTED] mailing list
