Re: Requesting assistance recovering RAID-5 array

["crowston.name" <kevin@crowston.name>]

I got the following error trying to run lsdrv: 

./lsdrv 
Traceback (most recent call last):
  File "./lsdrv", line 423, in <module>
    probe_block('/sys/block/'+x)
  File "./lsdrv", line 340, in probe_block
    blk.__dict__.update(extractvars(runx(['vol_id', '--export', '/dev/block/'+blk.dev])))
  File "./lsdrv", line 125, in runx
    out, err = sub.communicate()
  File "/usr/lib/python2.5/subprocess.py", line 667, in communicate
    return self._communicate(input)
  File "/usr/lib/python2.5/subprocess.py", line 1138, in _communicate
    rlist, wlist, xlist = select.select(read_set, write_set, [])
select.error: (4, 'Interrupted system call')




Kevin Crowston
206 Meadowbrook Dr.
Syracuse, NY 13210 USA
Phone: +1 (315) 464-0272
Fax: +1 (815) 550-2155

-----Original Message-----
From: Daniel Jones <dj@iowni.com>
Reply: Daniel Jones <dj@iowni.com>
Date: March 30, 2020 at 8:52:18 PM
To: antlists <antlists@youngman.org.uk>
Cc: linux-raid@vger.kernel.org <linux-raid@vger.kernel.org>
Subject:  Re: Requesting assistance recovering RAID-5 array

> Greetings Wol,
>  
> > Don't even THINK of --create until the experts have chimed in !!!
>  
> Yes, I have had impure thoughts, but fortunately (?) I've done nothing
> yet to intentionally write to the drives.
>  
> > If your drives are 1TB, I would *seriously* consider getting hold of a 4TB drive - they're  
> not expensive - to make a backup. And read up on overlays.
>  
> The array drives are 10TB each. Understand the concept of overlays in
> general (have used them in a container context) and have skimmed the
> wiki, but not yet acted.
>  
> > The lsdrv information is crucial - that recovers pretty much all the config information  
> that is available
>  
> Attached.
>  
> $ ./lsdrv
> PCI [pata_marvell] 02:00.0 IDE interface: Marvell Technology Group
> Ltd. 88SE6101/6102 single-port PATA133 interface (rev b2)
> └scsi 0:x:x:x [Empty]
> PCI [ahci] 00:1f.2 SATA controller: Intel Corporation 82801JI (ICH10
> Family) SATA AHCI Controller
> ├scsi 2:0:0:0 ATA M4-CT256M4SSD2 {0000000012050904283E}
> │└sda 238.47g [8:0] Partitioned (dos)
> │ ├sda1 500.00m [8:1] xfs {8ed274ce-4cf6-4804-88f8-0213c002a716}
> │ │└Mounted as /dev/sda1 @ /boot
> │ └sda2 237.99g [8:2] PV LVM2_member 237.92g used, 64.00m free
> {kn8lMS-0Cy8-xpsR-QRTk-CTRG-Eh1J-lmtfws}
> │ └VG centos_hulk 237.98g 64.00m free {P5MVrD-UMGG-0IO9-zFNq-8zd2-lycX-oYqe5L}  
> │ ├dm-2 185.92g [253:2] LV home xfs {39075ece-de0a-4ace-b291-cc22aff5a4b2}
> │ │└Mounted as /dev/mapper/centos_hulk-home @ /home
> │ ├dm-0 50.00g [253:0] LV root xfs {68ffae87-7b51-4392-b3b8-59a7aa13ea68}
> │ │└Mounted as /dev/mapper/centos_hulk-root @ /
> │ └dm-1 2.00g [253:1] LV swap swap {f2da9893-93f0-42a1-ba86-5f3b3a72cc9b}
> ├scsi 3:0:0:0 ATA WDC WD100EMAZ-00 {1DGVH01Z}
> │└sdb 9.10t [8:16] Partitioned (gpt)
> ├scsi 4:0:0:0 ATA WDC WD100EMAZ-00 {2YJ2XMPD}
> │└sdc 9.10t [8:32] MD raid5 (4) inactive 'hulk:0'
> {423d9a8e-636a-5f08-56ec-bd90282e478b}
> ├scsi 5:0:0:0 ATA WDC WD100EMAZ-00 {2YJDR8LD}
> │└sdd 9.10t [8:48] Partitioned (gpt)
> └scsi 6:0:0:0 ATA WDC WD100EMAZ-00 {JEHRKH2Z}
> └sde 9.10t [8:64] Partitioned (gpt)
>  
> Cheers,
> DJ
>  
> On Mon, Mar 30, 2020 at 6:24 PM antlists wrote:
> >
> > On 31/03/2020 01:04, Daniel Jones wrote:
> > > I am genuinely over my head at this point and unsure how to proceed.
> > > My logic tells me the best choice is to attempt a --create to try to
> > > rebuild the missing superblocks, but I'm not clear if I should try
> > > devices=4 (the true size of the array) or devices=3 (the size it was
> > > last operating in). I'm also not sure of what device order to use
> > > since I have likely scrambled /dev/sd[bcde] and am concerned about
> > > what happens when I bring the previously disable drive back into the
> > > array.
> >
> > Don't even THINK of --create until the experts have chimed in !!!
> >
> > https://raid.wiki.kernel.org/index.php/Linux_Raid#When_Things_Go_Wrogn  
> >
> > The lsdrv information is crucial - that recovers pretty much all the
> > config information that is available, and massively increases the
> > chances of a successful --create, if you do have to go down that route...
> >
> > If your drives are 1TB, I would *seriously* consider getting hold of a
> > 4TB drive - they're not expensive - to make a backup. And read up on
> > overlays.
> >
> > Hopefully we can recover your data without too much grief, but this will
> > all help.
> >
> > Cheers,
> > Wol
>