Tony Mountifield
2020-Sep-27 12:21 UTC
[CentOS] Using CentOS 7 to attempt recovery of failed disk
In article <E02FA554-9D6D-4E7D-8A78-5FBDE1DE939D at kicp.uchicago.edu>, Valeri Galtsev <galtsev at kicp.uchicago.edu> wrote:> > > > On Sep 26, 2020, at 8:05 AM, Jerry Geis <jerry.geis at gmail.com> wrote: > > > > I have a disk that is flagging errors, attempting to rescue the data. > > > > I tried dd first - if gets about 117G of 320G disk and stops incrementing > > the save image any more. > > did you try > > dd conv=noerror ??? > > this flag makes dd not stop on input error. Whatever is irrecoverable is irrecoverable, but this way you will get stuff > beyond failure point.You need conv=noerror,sync so that unreadable sectors get replaced by zeros instead of not being written out at all. Without sync, the filesystem geometry on the destination image will be wrong after the first error. You also need bs=4096 so that ONLY the bad sector(s) get zeroed, and not the surrounding ones. If you have, say, bs=1M, then you will get a megabyte of zeros if any block within that megabyte is bad. I'm speaking from recent experience! Cheers Tony -- Tony Mountifield Work: tony at softins.co.uk - http://www.softins.co.uk Play: tony at mountifield.org - http://tony.mountifield.org
Erick Perez - Quadrian Enterprises
2020-Sep-27 16:27 UTC
[CentOS] Using CentOS 7 to attempt recovery of failed disk
@tonymountifield Does this still hold true? https://superuser.com/a/1075837 On Sun, Sep 27, 2020 at 7:21 AM Tony Mountifield <tony at softins.co.uk> wrote:> In article <E02FA554-9D6D-4E7D-8A78-5FBDE1DE939D at kicp.uchicago.edu>, > Valeri Galtsev <galtsev at kicp.uchicago.edu> wrote: > > > > > > > On Sep 26, 2020, at 8:05 AM, Jerry Geis <jerry.geis at gmail.com> wrote: > > > > > > I have a disk that is flagging errors, attempting to rescue the data. > > > > > > I tried dd first - if gets about 117G of 320G disk and stops > incrementing > > > the save image any more. > > > > did you try > > > > dd conv=noerror ? > > > > this flag makes dd not stop on input error. Whatever is irrecoverable is > irrecoverable, but this way you will get stuff > > beyond failure point. > > You need conv=noerror,sync so that unreadable sectors get replaced by > zeros instead of not being written out at all. > Without sync, the filesystem geometry on the destination image will be > wrong after the first error. > > You also need bs=4096 so that ONLY the bad sector(s) get zeroed, and not > the surrounding ones. If you have, say, > bs=1M, then you will get a megabyte of zeros if any block within that > megabyte is bad. > > I'm speaking from recent experience! > > Cheers > Tony > > -- > Tony Mountifield > Work: tony at softins.co.uk - http://www.softins.co.uk > Play: tony at mountifield.org - http://tony.mountifield.org > _______________________________________________ > CentOS mailing list > CentOS at centos.org > https://lists.centos.org/mailman/listinfo/centos >-- --------------------- Erick Perez Quadrian Enterprises S.A. - Panama, Republica de Panama Skype chat: eaperezh WhatsApp IM: +507-6675-5083 ---------------------
Tony Mountifield
2020-Sep-28 16:37 UTC
[CentOS] Using CentOS 7 to attempt recovery of failed disk
In article <CACXMG+s-KcfMV_eFRqm12asaQ8rDABL123bD+ur8jwNVLyQXjw at mail.gmail.com>, Erick Perez - Quadrian Enterprises <eperez at quadrianweb.com> wrote:> @tonymountifield > Does this still hold true? > https://superuser.com/a/1075837It wouldn't surprise me. What I take away from those tests is that it is indeed important to use a bs= setting that corresponds to the disk physical block size, which is why I said to use bs=4096. When I used "conv=noerror,sync bs=4096" I got an image of the correct size. That seems to correspond with what is said in the comment you linked to. Cheers Tony> On Sun, Sep 27, 2020 at 7:21 AM Tony Mountifield <tony at softins.co.uk> wrote: > > > In article <E02FA554-9D6D-4E7D-8A78-5FBDE1DE939D at kicp.uchicago.edu>, > > Valeri Galtsev <galtsev at kicp.uchicago.edu> wrote: > > > > > > > > > > On Sep 26, 2020, at 8:05 AM, Jerry Geis <jerry.geis at gmail.com> wrote: > > > > > > > > I have a disk that is flagging errors, attempting to rescue the data. > > > > > > > > I tried dd first - if gets about 117G of 320G disk and stops > > incrementing > > > > the save image any more. > > > > > > did you try > > > > > > dd conv=noerror ??? > > > > > > this flag makes dd not stop on input error. Whatever is irrecoverable is > > irrecoverable, but this way you will get stuff > > > beyond failure point. > > > > You need conv=noerror,sync so that unreadable sectors get replaced by > > zeros instead of not being written out at all. > > Without sync, the filesystem geometry on the destination image will be > > wrong after the first error. > > > > You also need bs=4096 so that ONLY the bad sector(s) get zeroed, and not > > the surrounding ones. If you have, say, > > bs=1M, then you will get a megabyte of zeros if any block within that > > megabyte is bad. > > > > I'm speaking from recent experience! > > > > Cheers > > Tony > > > > -- > > Tony Mountifield > > Work: tony at softins.co.uk - http://www.softins.co.uk > > Play: tony at mountifield.org - http://tony.mountifield.org > > _______________________________________________ > > CentOS mailing list > > CentOS at centos.org > > https://lists.centos.org/mailman/listinfo/centos > > > > > -- > > --------------------- > Erick Perez > Quadrian Enterprises S.A. - Panama, Republica de Panama > Skype chat: eaperezh > WhatsApp IM: +507-6675-5083 > --------------------- > _______________________________________________ > CentOS mailing list > CentOS at centos.org > https://lists.centos.org/mailman/listinfo/centos >-- Tony Mountifield Work: tony at softins.co.uk - http://www.softins.co.uk Play: tony at mountifield.org - http://tony.mountifield.org