HELP no more Xwindows! libtermcap.so.2 cannot open shared object
in [Linux Help]
|
Prev: [FC9] Yum cannot update firefox
Next: HELP no more Xwindows! libtermcap.so.2 cannot open sharedobject
From: John Thompson on 7 Jan 2010 19:29 On 2010-01-06, Aragorn <aragorn(a)chatfactory.invalid> wrote: > /reiserfs/ (v3.6) is quite good and it's faster than /ext3/ but it lacks > a decent toolchain. XFS on the other hand is just as reliable > as /ext2/ or /ext3/ - if not more reliable, given that it's been the > default filesystem in IRIX since 1996 - but is a B+ tree filesystem > with extremely high performance, a complete toolchain (unlike reiserfs) > and a rather secure filesystem too, since it zeroes out damaged blocks. > It also allows for online snapshotting. It's acknowledged as an > industry standard filesystem. XFS was the default filesystem on my SGI "Indy" even before 1996. It's probably the most mature journaled filesystem in ther *nix world. -- -John (john(a)os2.dhs.org)
From: Nico Kadel-Garcia on 8 Jan 2010 00:13 On Jan 6, 6:27 pm, Aragorn <arag...(a)chatfactory.invalid> wrote: > On Wednesday 06 January 2010 16:35 in alt.os.linux, somebody identifying > > as Dan C wrote... > > On Wed, 06 Jan 2010 09:30:03 +0100, J.O. Aho wrote: > > >> IMHO ext3 suxx, just ext2 with journal added. > > > Exactly. Which makes it safe and stable. Good performance and > > stability. > > Well, that's in the eye of the beholder, Dan. ;-) > > > What more would you want from your file system? > > More performance and reliability, perhaps? ;-) Multiple B-plus trees, > for instance? ;-) > > > Please don't even mention the word "reiser"... > > /reiserfs/ (v3.6) is quite good and it's faster than /ext3/ but it lacks > a decent toolchain. XFS on the other hand is just as reliable > as /ext2/ or /ext3/ - if not more reliable, given that it's been the > default filesystem in IRIX since 1996 - but is a B+ tree filesystem > with extremely high performance, a complete toolchain (unlike reiserfs) > and a rather secure filesystem too, since it zeroes out damaged blocks. > It also allows for online snapshotting. It's acknowledged as an > industry standard filesystem. The following description is amazingly rude and insensitive, but heartfelt. Reiserfs treats your files like Hans Reiser treated his ex-wife. When something goes really wrong (such as divorce or a failing hard drive in a RAID set), it deletes them, pretends complete innocence, tries to hide the evidence, and wastes your time proving its guilt when we all know who did it. Like its author, it is technically brilliant but dangerously unstable. Unless your data actively enjoys S&M like Han's ex-wife did, and you would care to have your files risk the same fate, I can't recommend it. I had this sort of thing happen *every single time* I tried ReiserFS, with five different machines, two different operating systems (RHEL and SuSE). Since ext3 came out and addressed the "thousands of files in one directory" issue, there's simply no reason to use ReiserFS except possibly for high-volume NNTP or proxy servers where high performance matters and flesystem integrity is not so critical. > Another interesting project, albeit not quite stable yet, is the new > Btrfs (pronounced as "butter filesystem"), which has been submitted to > the Linux kernel by Oracle. It's still heavily in beta stage but it's > already included in the upstream sources from Linus & friends. It more > or less offers all the goodies that Sun's ZFS offers, i.e. an > integrated filesystem _and_ logical volume manager. It's an > interesting project, albeit not ready for prime time yet. I've been hearing good things: it seems to the equivalent of PostgreSQL versus MySQL: different details and different licensing, but a lot of the same overall functionality.
From: Jasen Betts on 8 Jan 2010 04:18 On 2010-01-06, Dan C <youmustbejoking(a)lan.invalid> wrote: > On Wed, 06 Jan 2010 09:08:16 -0600, Ohmster wrote: > >> "J.O. Aho" <user(a)example.net> wrote in news:7qj012Fqa7U1 >> @mid.individual.net: >> >>> IMHO ext3 suxx, just ext2 with journal added. >>> >>> -- >>> >>> //Aho >> >> I thought that the journal addition was a good thing. What do you like >> Aho? > > It is a good thing. EXT3 is the right choice. When I upgraded my hard drive I went with ext4 as I couldn't see any downsides, but I don't see any installers that offer it. anyone know why? --- news://freenews.netfront.net/ - complaints: news(a)netfront.net ---
From: J.O. Aho on 8 Jan 2010 05:58 Nico Kadel-Garcia wrote: > On Jan 6, 6:27 pm, Aragorn <arag...(a)chatfactory.invalid> wrote: >>> Please don't even mention the word "reiser"... >> /reiserfs/ (v3.6) is quite good and it's faster than /ext3/ but it lacks >> a decent toolchain. XFS on the other hand is just as reliable >> as /ext2/ or /ext3/ - if not more reliable, given that it's been the >> default filesystem in IRIX since 1996 - but is a B+ tree filesystem >> with extremely high performance, a complete toolchain (unlike reiserfs) >> and a rather secure filesystem too, since it zeroes out damaged blocks. >> It also allows for online snapshotting. It's acknowledged as an >> industry standard filesystem. > > The following description is amazingly rude and insensitive, but > heartfelt. > > Reiserfs treats your files like Hans Reiser treated his ex-wife. When > something goes really wrong (such as divorce or a failing hard drive > in a RAID set), it deletes them, pretends complete innocence, tries to > hide the evidence, and wastes your time proving its guilt when we all > know who did it. Like its author, it is technically brilliant but > dangerously unstable. Unless your data actively enjoys S&M like Han's > ex-wife did, and you would care to have your files risk the same fate, > I can't recommend it. > > I had this sort of thing happen *every single time* I tried ReiserFS, > with five different machines, two different operating systems (RHEL > and SuSE). Since ext3 came out and addressed the "thousands of files > in one directory" issue, there's simply no reason to use ReiserFS > except possibly for high-volume NNTP or proxy servers where high > performance matters and flesystem integrity is not so critical. I can say I never had any problems with reiserfs 3.6 (had a quite many bug fixes compared with earlier versions, which may be those you tried). On the other hand I have had bad file system problems with ext3, which wasn't repairable and lead to reinstalls of systems. Early Xfs for Linux I had quite many file losses, as much of the stuff was in ram and never pushed out to the hard drive when machine crashed. Jfs is my favorite, fast and by default goes to write only mode when something goes badly wrong, but needs more hand on after a system crash than resierfs (which I use it on unstable systems) and with both I get rid of the inode problem that you seem to have with ext2/3 with loads of small files and of course the laginesh of ext2/3 ain't there. -- //Aho
From: Bud on 8 Jan 2010 15:21
On 2010-01-08, Jasen Betts wrote: > > When I upgraded my hard drive I went with ext4 as I couldn't see any > downsides, but I don't see any installers that offer it. > > anyone know why? No, but Mandriva offers it as well as Slackware. As to resierfs and resierfs4 I've found no problems with either. Luck or they just work? -- Bud |