On Sun, 2006-02-12 at 21:02 +0100, Ingo wrote: > When reading data from OS/2 or Linux , JFS beheaves very co-operative: > existing EA's from OS/2 are not touched by Linux, even when jfs_fsck is run > existing permission and ownership-bits for Linux are not touched by OS/2! > the EA's from OS/2 are not copied under Linux > this is a great way to strip all EA's if wanted, like when copying to > NFS via TCP/IP > the permissions of Linux are not copied under OS/2 > this is one more reason to apply Dave's patch for access. > So data exchange is working perfectly now.
By the way, if you did want to access the os2 EA's from linux, you should be able to get to them through libattr (getfattr & co.) using the "os2." prefix. > Is there any other place besides souceforge where more detailled information > on JFS > can be found? Under OS/2 for instance you are able to tune the timing > parameters for the > caching like lazywrite, synctime, maxage, bufferidle. Does also under Linux > JFS have some > configuration file? In Linux, jfs doesn't manage most of this, and relies on the kernel's vm to handle it. You may want to look into tuning the values under /proc/sys/vm. > Does JFS accept the option 'sync' for mounting? Yes > The background for this question is that occasionally, when copying thousands > of small files > (2000 directories, 20000 files, each less than 1 MB) by using the Konqueror > under KDE some > files got lost. Linux jfs_fschk (umounted before) did not report any > anomalities, > but OS/2's 'chkdsk /f:2' gave several lost handles and many files in > /lost+found. There is a bug in Linux, that jfs does not maintain the size of the directory correctly on legacy partitions, which makes os/2's fsck unhappy. I tried fixing this here: http://bugzilla.kernel.org/show_bug.cgi?id=5549 but it didn't seem to fix it. I need to look into it some more. > Only the target data were affected by that loss, all original data were 100% > ok and > after running chkdsk /f:2 the filesystem was absolutely clean. > > Under OS/2, I was able to solve a similar beheaviour by playing with the > cache settings: > 8 MB cache only, lazywrite, synctime=5s, maxage=20sec, bufferidle=4sec > > However it also could be that this is one incompatibility in KDE when > handling so many > files in one run by just drag&drop of the directory? I don't think it's really related to kde. I think anything that creates enough entries in a directory to affect it's size will cause this bug. > With best regards, > Ingo > > -- David Kleikamp IBM Linux Technology Center ------------------------------------------------------- This SF.net email is sponsored by: Splunk Inc. Do you grep through log files for problems? Stop! Download the new AJAX search engine that makes searching your log files as easy as surfing the web. DOWNLOAD SPLUNK! http://sel.as-us.falkag.net/sel?cmd=lnk&kid=103432&bid=230486&dat=121642 _______________________________________________ Jfs-discussion mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/jfs-discussion
