ISW dmraid-rc15 fails to activate raid0 due to incorrect cylinder calculation
Labun, Marcin
Marcin.Labun at intel.com
Tue Apr 21 12:44:21 UTC 2009
Hello Giuseppe,
ISW metadata blocks_per_member (dev->vol.map.blocks_per_member) shall contain journaling space on the RAID member device, while r->sectors shall contain data space only.
During RAID device configuration in metadata the journaling space is added to blocks_per_member:
map->blocks_per_member = (map->blocks_per_strip * map->num_data_stripes) * div + RAID_DS_JOURNAL;
Could you please provide ISW native metadata output from dmraid -n and device mapper mapping (dmsetup table and status) and dmraid array configuration (dmraid -C and dmraid -s)?
Do you see "%s: zero sectors on %s"?
If not, possibly there is an overflow and the reason might be in blocks_per_member configuration.
Regards,
Marcin
> -----Original Message-----
> From: ataraid-list-bounces at redhat.com [mailto:ataraid-list-
> bounces at redhat.com] On Behalf Of Giuseppe Iuculano
> Sent: Sunday, April 19, 2009 12:42 PM
> To: ATARAID (eg, Promise Fasttrak, Highpoint 370) related discussions
> Cc: 524637-forwarded at bugs.debian.org
> Subject: Re: ISW dmraid-rc15 fails to activate raid0 due to incorrect
> cylinder calculation
>
> Hi,
>
> Danny ha scritto:
> > Upon many attempts and comparisons to rc14 I discovered that
> > dmraid-rc15 calculates the number of cylinders incorrectly..
>
> The relevant code in isw.c is:
>
> if ((r->sectors = dev->vol.map.blocks_per_member - RAID_DS_JOURNAL))
>
> But in dmraid rc14 was:
>
> if ((r->sectors = dev->vol.map.blocks_per_member))
>
>
> What is the purpose of this change?
>
> Cheers,
> Giuseppe.
More information about the Ataraid-list
mailing list