Indeed, after creating an empty gpt empty partition table with an ubuntu bootable usb key, I was able to add the disk to the mirror. the mirror appeared to work fine after resilvering. The situation was then that I had a mirror rpool with an original 2TB disk and a resilvered 4 TB disk of which only 2 TB are used. Then I tried to reboot on the 2 TB disk and, during booting, I got  lot of messages about read errors on a disk at sector 7.814.037.168.

Warning: /pci@0,0/pci-ide@1f,2/ide@1f,2/ide@1/cmdk@O,O (Disk3):

                Error for command "read sector' Error level: Fatal

                Requested block 7 814 037 168, Error block 7 814 037 168

                Sense key: ID not found

                Vendor 'Gen-ATA' error code: 0x5



If I check the gpt table with format/verify, I get

format> verify
Reading the primary EFI GPT label failed.  Using backup label.
Use the 'backup' command to restore the primary label.

Volume name = <        >
ascii name  = <ST4000VN008-2DR166= ZGY5QJC9-            ZGY5QJC9-0001-3.64TB>
bytes/sector    =  512
sectors = 7813951487
accessible sectors = 7813951454
Part      Tag    Flag     First Sector          Size          Last Sector
  0        usr    wm               256         3.64TB 7813935070
  1 unassigned    wm                 0            0 0
  2 unassigned    wm                 0            0 0
  3 unassigned    wm                 0            0 0
  4 unassigned    wm                 0            0 0
  5 unassigned    wm                 0            0 0
  6 unassigned    wm                 0            0 0
  8   reserved    wm        7813935071         8.00MB 7813951454

format>
Thus, the faulty sector is beyond the end of the disk.

So why is  the GPT label unreadable in the first place and why does openindiana (2016.04) try to read so far on the disk ?

Here are the properties of the pool ?

root@spitfire:~# zpool get all
NAME   PROPERTY VALUE                          SOURCE
rpool  size 1,81T                          -
rpool  capacity 86%                            -
rpool  altroot -                              default
rpool  health DEGRADED                       -
rpool  guid                       040000648509983196 default
rpool  version -                              default
rpool  bootfs                 rpool/ROOT/openindiana-2       local
rpool  delegation on                             default
rpool  autoreplace off                            default
rpool  cachefile -                              default
rpool  failmode wait                           default
rpool  listsnapshots off                            default
rpool */autoexpand on                        local/*
rpool  dedupditto 0                              default
rpool  dedupratio 1.00x                          -
rpool  free 246G                           -
rpool  allocated 1,57T                          -
rpool  readonly off                            -
rpool  comment -                              default
rpool  expandsize -                              -
rpool  freeing 0                              default
rpool  fragmentation 41%                            -
rpool  leaked 0                              default
rpool  bootsize -                              default
rpool  feature@async_destroy  enabled                      local
rpool  feature@empty_bpobj    active                         local
rpool  feature@lz4_compre      active local
rpool  feature@multi_vdev_crash_dump enabled                        local
rpool  feature@spacemap_histogram active                         local
rpool  feature@enabled_txg active                         local
rpool  feature@hole_birth active                         local
rpool  feature@extensible_dataset enabled                        local
rpool  feature@embedded_data active                         local
rpool  feature@bookmarks enabled                        local
rpool  feature@filesystem_limits enabled                        local
rpool  feature@large_blocks enabled                        local
rpool  feature@sha512 enabled                        local
rpool  feature@skein enabled                        local
rpool  feature@edonr enabled                        local
root@spitfire:~#

I tried to set autoexpand to off and rebooted. The error messages were still there and the system could not finish booting. the, I removed physicalle the 4TB disk and the system could boot but without the mirror.

Can anybody help me with this ?

Thanks in advance


Marc


On 6/04/20 00:51, Reginald Beardsley via openindiana-discuss wrote:
  I had to use Linux (CentOS 6.7) to write a Sun label and create a gpt 
partition table with parted before either Hipster 2017.10 or Solaris 10 would 
recognize a 3 TB disk on my Z400s.

On Sol 10 u8 "format -e" dumped core. On Hipster 2017.10 it simply did not see the disk. 
But once I labeled it with Linux I had no further trouble. "format -e" happily made a 
proper Sun label on Sol 10. I don't think that there was anything special about CentOS 6.7. It 
happened to be the Linux distro I was using at the time. Since I've switched to Debian 9.3. I'm 
pretty sure I used Debian to set up the disk when I added a mirror for the 3 TB disk, but I didn't 
find any notes about that. So I have to presume I just stuck the drive in and followed my notes 
from before when I was replacing a failed 512 sector 3 TB disk with a 2k sector disk.

I'm running three 2 TB disks in RAIDZ1 for the export pool and a 3 way mirror 
for rpool using a pair of slices both for Hipster 2017.10 and Sol 10. I have 
the Sol 10 box offline on an isolated network. I prefer twm for programming 
work and that is far more work to set up on Hipster than I care to endure. So 
Hipster is my Internet box and Sol 10 is my serious work box. At one time I ran 
twm on one screen and CDE on the other to keep 3rd party stuff happy.

The Sol 10 box has a 2 way 3 TB mirror scratch space. Probably not needed now, 
but I used to do seismic processing which needs lots of scratch space. I've 
also got an N40L with RAIDZ2 on four 2 TB disks with a 4 way mirror for rpool 
for backups.

Good luck,
Reg
      On Sunday, April 5, 2020, 07:03:39 PM CDT, Marc 
Lobelle<[email protected]>  wrote:
Hello,

I'm trying to install a disk (seagate Ironwolf 4TB type ST4000VN008) on
a system running hipster 2016.04. The computer is based on an ASUS P8B-M
mother board with a Xeon processor. The sata interface works with an
older 2TB disk

(I know it is an old version of hipster, but my plan is first to replace
the 2 mirror disks by bigger ones, then update the OS.

The problem is that hipster does not identify the disk. The disk is new.
I enev tried 2 different ones.

# format
Searching for disks...done


AVAILABLE DISK SELECTIONS:
         0. c3d0 <Unknown-Unknown-0001 cyl 60796 alt 2 hd 255 sec 252>
|                                   an old one
/pci@0,0/pci-ide@1f,2/ide@0/cmdk@0,0
                                        |   the 2 mirror disks
         1. c6d0 <drive type
unknown><<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<< here
|                                   a new one
/pci@0,0/pci-ide@1f,2/ide@1/cmdk@0,0 |
         2. c6d1 <EAGET S606 SSD
120GB=G32350M011540-G32350M011540-0001-111.79GB>            A small ssd
used as cache by zfs
            /pci@0,0/pci-ide@1f,2/ide@1/cmdk@1,0
Specify disk (enter its number): ^C
#
Anybody has an idea where to look for ?

Thanks


Marc


_______________________________________________
openindiana-discuss mailing list
[email protected]
https://openindiana.org/mailman/listinfo/openindiana-discuss
_______________________________________________
openindiana-discuss mailing list
[email protected]
https://openindiana.org/mailman/listinfo/openindiana-discuss

_______________________________________________
openindiana-discuss mailing list
[email protected]
https://openindiana.org/mailman/listinfo/openindiana-discuss

Reply via email to