On Fri, Jul 10, 2015 at 05:57:15PM +0000, Vijay Sankar wrote: > Quoting Otto Moerbeek <[email protected]>: > > >On Fri, Jul 10, 2015 at 04:04:04PM +0000, Vijay Sankar wrote: > > > >>My objective for this weekend was to follow the new dpb and build ports > >>without using sudo. So I was hoping to upgrade to the latest snapshot on a > >>system that I use for tests. > >> > >>The test system has a 2TB drive and it had two 300GB partitions in it for > >>ports and vm; and a 120GB SSD for the OS and used to look as follows: > >> > >>Filesystem Size Used Avail Capacity Mounted on > >>/dev/sd1a 1005M 55.0M 900M 6% / > >>/dev/sd1k 64.5G 20.9G 40.3G 34% /home > >>/dev/sd1d 3.9G 10.0K 3.7G 0% /tmp > >>/dev/sd1f 2.0G 966M 946M 51% /usr > >>/dev/sd1g 1005M 191M 764M 20% /usr/X11R6 > >>/dev/sd1h 9.8G 2.9G 6.5G 31% /usr/local > >>/dev/sd1j 2.0G 2.0K 1.9G 0% /usr/obj > >>/dev/sd1i 2.0G 827M 1.1G 43% /usr/src > >>/dev/sd1e 13.5G 26.5M 12.8G 0% /var > >>/dev/sd0h 298G 176G 107G 62% /ports > >>/dev/sd0f 298G 19.6G 263G 7% /vm > >> > >>My /etc/fstab was > >> > >>4f0cd8b5e7fd8f6a.b none swap sw > >>4f0cd8b5e7fd8f6a.a / ffs rw 1 1 > >>4f0cd8b5e7fd8f6a.k /home ffs rw,nodev,nosuid 1 2 > >>4f0cd8b5e7fd8f6a.d /tmp ffs rw,nodev,nosuid 1 2 > >>4f0cd8b5e7fd8f6a.f /usr ffs rw,nodev 1 2 > >>4f0cd8b5e7fd8f6a.g /usr/X11R6 ffs rw,nodev 1 2 > >>4f0cd8b5e7fd8f6a.h /usr/local ffs rw,nodev 1 2 > >>4f0cd8b5e7fd8f6a.j /usr/obj ffs rw,nodev,nosuid 1 2 > >>4f0cd8b5e7fd8f6a.i /usr/src ffs rw,nodev,nosuid 1 2 > >>4f0cd8b5e7fd8f6a.e /var ffs rw,nodev,nosuid 1 2 > >>4d43e3389228e319.h /ports ffs rw,nodev,nosuid 1 2 > >>4d43e3389228e319.f /vm ffs rw,nodev,nosuid 1 2 > >> > >>I am not sure what happened -- but when I rebooted the system this morning > >>/ports and /vm would not mount; so I commented out the last two lines in > >>/etc/fstab and rebooted. After reboot disklabel seems to have changed > >>completely and it currently looks like this: > >> > >># disklabel sd0 > >># /dev/rsd0c: > >>type: SCSI > >>disk: SCSI disk > >>label: ST2000DM001-1CH1 > >>duid: 0000000000000000 > >>flags: > >>bytes/sector: 512 > >>sectors/track: 63 > >>tracks/cylinder: 255 > >>sectors/cylinder: 16065 > >>cylinders: 503 > >>total sectors: 8089950 > >>boundstart: 0 > >>boundend: 8089950 > >>drivedata: 0 > >> > >>16 partitions: > >># size offset fstype [fsize bsize cpg] > >> c: 8089950 0 unused > >> > >> > >>Is there any way fix the disklabel or is this an error that is impossible to > >>recover from? duid used to show up as 4d43e3389228e319 and not > >>0000000000000000. > >> > >>Please let me know if you have any suggestions. > > > >Get your old label from /var/backups and try to restore it with > >disklabel -R. You don't tell what your platform is, it might be that > >you also need to do fdisk work first to restore the mbr partition > >table. > > > >But of course, it is also interesting to know what happened to you > >disk. But since you do not tell us what you did you are on your own > >here. > > > > -Otto > > Thank you very much. I am running an older snapshot OpenBSD 5.7 -current as > of Mar 19, 2015. I thought of -R with disklabel but since the drive seems to > show itself as a 3950MB drive instead of a 2TB drive, I was not sure how to > do this. > > The problem truly is I am not sure what I did to cause all this problem!!! > The sequence of actions were as follows. Since I had not looked at this box > for a while I was just logging in to look at where I had kept everything. I > did a cd /ports/packages/amd64/all and got an input error when I tried to > edit a file. So I did a shutdown -h now; opened the 3.5" and 2.5" hotswap > drive bays and pulled both drives out and pushed them back in. Powered the > system on at which point I was dropped into the shell because /vm and /ports > had errors. So I tried to do a fsck_ffs and that failed. At that point I > looked at disklabel and noticed that the duid was gone. fdisk sd0 does not > show anything other than: > > # fdisk sd0 > Disk: sd0 geometry: 503/255/63 [8089950 Sectors] > > I tried the disklabel -R as you suggested; > > # disklabel -R sd0 disklabel.sd0.current > disklabel: partition a: partition extends past end of unit > disklabel: partition c: partition extends past end of unit > disklabel: partition d: offset past end of unit > disklabel: partition d: partition extends past end of unit > disklabel: partition e: offset past end of unit > disklabel: partition e: partition extends past end of unit > disklabel: partition f: offset past end of unit > disklabel: partition f: partition extends past end of unit > disklabel: partition g: offset past end of unit > disklabel: partition g: partition extends past end of unit > disklabel: partition h: offset past end of unit > disklabel: partition h: partition extends past end of unit > disklabel: partition i: offset past end of unit > disklabel: partition i: partition extends past end of unit > > Also tried > # fdisk -i sd0 > Do you wish to write new MBR and partition table? [n] y > Writing MBR at offset 0. > fdisk: error writing MBR: Input/output error > > Not sure whether there is any other option but Thanks very much for the help > and advice.
Try checking the connectors in the hot-swap bay and use compressed air to remove the dust. A couple of times I have experienced UDMA errors when the drive was fine, but the connections were not. In my case the causes were dust or SATA cables leaning on a side. Maybe the transfers get corrupted, so the OS reads wrong geometry? Mateusz -- Mateusz Lenik PGP B865 E86A D36C 11A5 C1F8 C1D9 AAD4 CEC9 6B94 92C4
