> On Mar 9, 2020, at 1:51 PM, Bhupesh Sharma <[email protected]> wrote:
>
> Hi John,
>
> On Mon, Mar 9, 2020 at 10:38 PM John Donnelly
> <[email protected]> wrote:
>>
>> Hi kexec/kdump team.
>>
>> I am not sure if this is the appropriate dlist to ask . If not , I
>> apologize .
>>
>>
>> I am having difficulties mounting a iSCSI target under kdump .
>
> We have had several known issues with iSCSI targets under kdump which
> have been fixed, so just wanted to check which kexec-tools version you
> are using: upstream or fedora?
>
> Can you please share the output of:
> $ kexec -v
# kexec -v
kexec-tools 2.0.15
/usr/sbin/kexec
# rpm -qf `which kexec `
kexec-tools-2.0.15-33.0.9.el7.x86_64
>
>> The target is discovered ,
>>
>> [ 154.118729] sd 2:0:0:0: [sda] Attached SCSI disk
>> [ OK ] Found device ZFS_Storage_7350 4.
>> Starting File System Check on /dev/...a-5ef4-4838-b5e7-dff852dfc673...
>> [ OK ] Started File System Check on /dev/d...46a-5ef4-4838-b5e7-dff852dfc673.
>> [ 275.766578] dracut-initqueue[420]: Warning: dracut-initqueue timeout -
>> startinging
>>
>>
>> The fsck step never finishes, and dracut timeout eventually drops into the
>> rescue shell.
>>
>>
>> I can see it is attempting to fsck root from the UUID :
>>
>>
>> # blkid | grep b5
>> /dev/sda4: UUID="fb01846a-5ef4-4838-b5e7-dff852dfc673" TYPE="xfs"
>> PARTUUID="f8831f2d-b2c6-4b27-97db-0513e4d3fe42"
>> 3:02
>> " /dev/sda4 30G 3.5G 27G 12% /
>>
>>
>> And I mount root manually and look around.
>>
>> # kdump:/# mkdir /mnt
>> # kdump:/# mount /dev/sda4 /mnt
>> [ 542.856035] SGI XFS with ACLs, security attributes, realtime, scrub,
>> repair, no debug enabled
>> [ 542.885016] XFS (sda4): Mounting V4 Filesystem
>> [ 542.938507] XFS (sda4): Starting recovery (logdev: internal)
>> [ 542.994720] XFS (sda4): Ending recovery (logdev: internal)
>> [ 543.018918] xfs filesystem being mounted at /mnt supports timestamps until
>> 2038 (0x7fffffff)
>> kdump:/#
>> kdump:/#
>> kdump:/# chroot /mnt/ /usr/bin/bash
>> bash-4.2#
>>
>> Now my iSCSI target is mounted as /mnt
>>
>>
>> Is there a way start dracut so it stops BEFORE the fsck step ? Not after
>> it fails ?
>
> I think you can try using 'rd.break' dracut option:
>
>
> rd.break={cmdline|pre-udev|pre-trigger|initqueue|pre-mount|mount|pre-pivot|cleanup}
> drop to a shell on defined breakpoint
>
Thanks for the hint ! I will try these.
> you can specify the same in the kdump bootargs, by modifying a system
> configuration file (for e.g. its '/etc/sysconfig/kdump' on fedora/rhel
> systems).
>
> For example, you can use 'rd.break=cmdline' to drop to the dracut
> shell to try and see if you can stop it before the 'fsck' step.
>
> Thanks,
> Bhupesh
>
_______________________________________________
kexec mailing list
[email protected]
http://lists.infradead.org/mailman/listinfo/kexec
