-------- Forwarded Message -------- Subject: Re: [ntfs-3g-devel] mkntfs: The partition start sector specified for /dev/sdc5 and the automatically determined value is too large. It has been set to 0. Date: Sun, 24 May 2015 08:53:07 +0200 From: Jean-Pierre André <jean-pierre.an...@wanadoo.fr> To: Martin Mokrejs <mmokr...@fold.natur.cuni.cz>
Martin Mokrejs wrote: > Hi Jean-Pierre, > > thank you for your quick answer. Comments below. > > Jean-Pierre André wrote: >> Hi, >> >> Martin Mokrejs wrote: >> >> > Hi, >> > I am configuring a Linux under VirtualBox-5.0.0-rc3 and one of the >> > partition will be NTFS. The disk is raw disk, on host exposed as >> > /dev/sdc, on guest appears as /dev/sda. >> > >> > It seems detection of virtual drive geometry fails or mkntfs is >> broken. >> > Seems I can provide mkntfs with cyl/head/sect counts but what is the >> > cause? >> > >> > It seems very dangerous to blindly set starting sector to zero and >> > starting zero-ing my drive. >> >> The starting sector of an NTFS partition is only used for booting >> Windows, and it is a 32-bit field which cannot hold 5860533134. Please read 4357349376 (initial sector of sdc5) >> There is no problem with a null starting sector if this not a >> Windows system partition. > > Well, in my case my aim was to use it only for data, so fine with me, but, > would you please make the message from mkntfs clearer? Also, could it > include > the actually obtained value, so 5860533134 in my case? And explain the > 32-bit issue? > > >> Likewise you cannot encode 5860533134 into C/H/S values (even >> manually !), and they are only used for booting Windows. >> >> Both are values to be stored in the boot sector of the NTFS >> partition (not the boot sector of the drive). [...] >> This is unusual, so far all the 3TB drives I have heard of >> have 4K sectors. As a consequence in your case the number >> of sectors does not fit into a 32-bit word. > > Hmm, this is Lacie D2 QUADRA 3TB drive, pretty expensive one with eSATA, > FireWire, USB3.0. So I should look for a firmware flash? ;-) This can also result from the hardware emulating a 4K-byte sector on 512-byte sectors... or VirtualBox doing the emulation. In both cases writing a 512-byte sector implies reading a 4K sector, updating 512 bytes and writing the 4K sector, which is inefficient. You should ask the vendor. >> You should have dumped the beginning of /dev/sdc5 to locate >> an NTFS partition with the fields related to the geometry of >> the drive left unfilled. [...] > Aha. How? Via dd with offset=4357349376-1 ? Either : dd if=/dev/sdc5 bs=512 count=1 | od -t x1 or (provided 4357349376 does not get truncated) : dd if=/dev/sdc skip=4357349376 bs=512 count=1 | od -t x1 or : dd if=/dev/sdc skip=544668672 bs=4096 count=1 | od -t x1 | head -n 32 > >> >> Main question : did you format /dev/sdc5 and is it usable ? > > I did not let it continue, I was so scared to zap my raw drive ... ;-) > I will first resolve with you what is in the NTFS boot block .;-) Oh, and if you do not use option -f it will take hours with no value added. Regards Jean-Pierre ------------------------------------------------------------------------------ One dashboard for servers and applications across Physical-Virtual-Cloud Widest out-of-the-box monitoring support with 50+ applications Performance metrics, stats and reports that give you Actionable Insights Deep dive visibility with transaction tracing using APM Insight. http://ad.doubleclick.net/ddm/clk/290420510;117567292;y _______________________________________________ ntfs-3g-devel mailing list ntfs-3g-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/ntfs-3g-devel
