On Tue, Jan 31, 2006 at 02:28:35PM +0100, Szakacsits Szabolcs wrote: > > Hi, > > On Tue, 31 Jan 2006, Yuval Fledel wrote: > > > You might be interested in that. > > Thanks Yuval for forwarding it from the Debian bug tracking system! > > > Package: ntfsprogs > > Version: 1.12.1-1 > > > > Since upgrading to the latest version, evms has become very slow on launch. > > I guess, since upgrading evms, not ntfsprogs, right? I didn't notice > ntfsresize would have become significantly slower ever.
ntfsprogs was new; evms didn't change. Reviewing the logs, I see this was a new install, not an upgrade. mindi in Debian changed to depend on ntfsprogs, and I guess evms uses them if they are available. > > > Examining the debug logs, I see > > Jan 30 22:01:06 wheat _8_ Engine: do_fork_and_execvp: execvp => > > ntfsresize -i -f /dev/evms/hda7 > > Jan 30 22:01:46 wheat _6_ NTFS: get_min_fs_size: ntfsresize completed > > with exit code 0 > > > > So it looks as if ntfsresize is taking 40 seconds to execute. I've > > verified this from the command line, and it remains true if I use > > /dev/hda7 rather than /dev/evms/hda7. > > > > It appears that a lot of checks are taking place, so maybe this is a > > good thing. But maybe not, hence this report. > > The checks are intentional and it can take such long time if the size of > the $MFT is big enough or you have IDE DMA disabled (see, hdpram > /dev/hda). Is hdpram a program? I can't find it, or any closely related spelling, in the Debian archives. > > > If the behavior is intended, might a mode that is less careful be > > possible? > > For what reason? So ntfsresize could run faster for informational queries. I'm not sure this is possible for ntfsresize or desirable in the context of evms, which (I think, in the latest versions) may do some write operations on ntfs partitions (such as resizing). I've received no feedback from the evms list, so far. > > Best regards, > > Szaka Thanks for the tools. Ross -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
