* On Wed Dec 05, 2001 at 01:26:39AM +1100, Matt Hyne ([EMAIL PROTECTED]) wrote: > At Tuesday, 4/12/2001 10:13 PM (+1100), Tony Green wrote: > >* This one time, at band camp, Craige McWhirter said: > >> On Tue, 2001-12-04 at 09:04, Xiaolu Zhang wrote: > >> > Last night due to the power cut the redhat linux 6.2 webserver was switch > >> > off without showdown. Now I got a error saying that one of the patition " > >> > has system file error , please run fsck manually " . what should I do ? > >> > >> While not an immediate solution but something well worth looking at if you > >> are at the stage of building your kernels, is the XFS journalling file > >> systems. Implementing a journalling file system will make recovery from > >> unexpected outages a whole lot more pleasant. > >> > > > >And since you are running a redhat 6.2 system (unless you have kept VERY > >up to date with patches) - an update may well be in order. > > > >(I was going to suggest this in my first post - but felt a 'this > >wouldn't happen if you had [xfs|ext3|reiser]' wouldn't go down well) > > It was my impression that none of the Journalling fs are recommended for production >environments just yet. > > If no so - correct me - I am interested to know. >
Depends on your definition of ready really. I use 2.2.x on most of my servers (simply because they don't need the 2.4 kernel tree functionality), so journalling isn't a viable option (without seriously patching the kernel). However, I have been using reiserfs for quite some time now (over a year) and haven't had a problem with it. I've been using it on my own servers and laptop. I'm also in the process of test ext3 - which is apparently safer than reiser. AFAIK Redhat is shipping with the 2.4 kernel tree now and MAY (I don't know if RH are but SuSE are) giving journalling options during installation. It all depends on your requirements. In the environment I am in, Linux boxes are used as satalites to larger machines and/or network storage - therefore priority is on ensuring quick reboots. If we totally loose a machine, we kickstart it from scratch and regain its former position. Therefore linux journalling is more of a viable option here. If you are using your linux machines for more business critical applications where you are storing data which can't be replaced on them, I'd stick with ext2 and good backups. As with everything in the Linux world, the exact solution really depends on your individual beliefs, requirements and expertise. HTH Greeno -- Tony Green <[EMAIL PROTECTED]> Tel : +61-(0)2-9500-9996 Mobile : +61-(0)4-2521-9996 GnuPG Key : 1024D/B5657C8B Key fingerprint = 9ED8 59CC C161 B857 462E 51E6 7DFB 465B B565 7C8B Dyslexic man walks into a bra -- SLUG - Sydney Linux User Group Mailing List - http://slug.org.au/ More Info: http://lists.slug.org.au/listinfo/slug
