[zfs-discuss] cannot replace c10t0d0 with c10t0d0: device is too small
In 2007 I bought 6 WD1600JS 160GB sata disks and used 4 to create a raidz storage pool and then shelved the other two for spares. One of the disks failed last night so I shut down the server and replaced it with a spare. When I tried to zpool replace the disk I get: zpool replace tank c10t0d0 cannot replace c10t0d0 with c10t0d0: device is too small The 4 original disk partition tables look like this: Current partition table (original): Total disk sectors available: 312560317 + 16384 (reserved sectors) Part TagFlag First Sector Size Last Sector 0usrwm34 149.04GB 312560350 1 unassignedwm 0 0 0 2 unassignedwm 0 0 0 3 unassignedwm 0 0 0 4 unassignedwm 0 0 0 5 unassignedwm 0 0 0 6 unassignedwm 0 0 0 8 reservedwm 3125603518.00MB 312576734 Spare disk partition table looks like this: Current partition table (original): Total disk sectors available: 312483549 + 16384 (reserved sectors) Part TagFlag First Sector Size Last Sector 0usrwm34 149.00GB 312483582 1 unassignedwm 0 0 0 2 unassignedwm 0 0 0 3 unassignedwm 0 0 0 4 unassignedwm 0 0 0 5 unassignedwm 0 0 0 6 unassignedwm 0 0 0 8 reservedwm 3124835838.00MB 312499966 So it seems that two of the disks are slightly different models and are about 40mb smaller then the original disks. I know I can just add a larger disk but I would rather user the hardware I have if possible. 1) Is there anyway to replace the failed disk with one of the spares? 2) Can I recreate the zpool using 3 of the original disks and one of the slightly smaller spares? Will zpool/zfs adjust its size to the smaller disk? 3) If #2 is possible would I still be able to use the last still shelved disk as a spare? If #2 is possible I would probably recreate the zpool as raidz2 instead of the current raidz1. Any info/comments would be greatly appreciated. Robert -- Robert Hartzell b...@rwhartzell.net RwHartzell.Net, Inc. ___ zfs-discuss mailing list zfs-discuss@opensolaris.org http://mail.opensolaris.org/mailman/listinfo/zfs-discuss
Re: [zfs-discuss] cannot replace c10t0d0 with c10t0d0: device is too small
On Mar 4, 2011, at 10:01 AM, Tim Cook wrote: On Fri, Mar 4, 2011 at 10:22 AM, Robert Hartzell b...@rwhartzell.net wrote: In 2007 I bought 6 WD1600JS 160GB sata disks and used 4 to create a raidz storage pool and then shelved the other two for spares. One of the disks failed last night so I shut down the server and replaced it with a spare. When I tried to zpool replace the disk I get: zpool replace tank c10t0d0 cannot replace c10t0d0 with c10t0d0: device is too small The 4 original disk partition tables look like this: Current partition table (original): Total disk sectors available: 312560317 + 16384 (reserved sectors) Part TagFlag First Sector Size Last Sector 0usrwm34 149.04GB 312560350 1 unassignedwm 0 0 0 2 unassignedwm 0 0 0 3 unassignedwm 0 0 0 4 unassignedwm 0 0 0 5 unassignedwm 0 0 0 6 unassignedwm 0 0 0 8 reservedwm 3125603518.00MB 312576734 Spare disk partition table looks like this: Current partition table (original): Total disk sectors available: 312483549 + 16384 (reserved sectors) Part TagFlag First Sector Size Last Sector 0usrwm34 149.00GB 312483582 1 unassignedwm 0 0 0 2 unassignedwm 0 0 0 3 unassignedwm 0 0 0 4 unassignedwm 0 0 0 5 unassignedwm 0 0 0 6 unassignedwm 0 0 0 8 reservedwm 3124835838.00MB 312499966 So it seems that two of the disks are slightly different models and are about 40mb smaller then the original disks. I know I can just add a larger disk but I would rather user the hardware I have if possible. 1) Is there anyway to replace the failed disk with one of the spares? 2) Can I recreate the zpool using 3 of the original disks and one of the slightly smaller spares? Will zpool/zfs adjust its size to the smaller disk? 3) If #2 is possible would I still be able to use the last still shelved disk as a spare? If #2 is possible I would probably recreate the zpool as raidz2 instead of the current raidz1. Any info/comments would be greatly appreciated. Robert You cannot. That's why I suggested two years ago that they chop off 1% from the end of the disk at install time to equalize drive sizes. That way you you wouldn't run into this problem trying to replace disks from a different vendor or different batch. The response was that Sun makes sure all drives are exactly the same size (although I do recall someone on this forum having this issue with Sun OEM disks as well). It's ridiculous they don't take into account the slight differences in drive sizes from vendor to vendor. Forcing you to single-source your disks is a bad habit to get into IMO. --Tim Well that sucks... So I guess the only option is to replace the disk with a larger one? Or are you saying thats not possible either? I can upgrade to larger disks but then there is no guarantee that I can even buy 4 identical disks off the shelf at any one time. Thanks for the info -- Robert Hartzell b...@rwhartzell.net RwHartzell.Net, Inc. ___ zfs-discuss mailing list zfs-discuss@opensolaris.org http://mail.opensolaris.org/mailman/listinfo/zfs-discuss
Re: [zfs-discuss] cannot replace c10t0d0 with c10t0d0: device is too small
On Mar 4, 2011, at 11:46 AM, Cindy Swearingen wrote: Robert, Which Solaris release is this? Thanks, Cindy Solaris 11 express 2010.11 -- Robert Hartzell b...@rwhartzell.net RwHartzell.Net, Inc. ___ zfs-discuss mailing list zfs-discuss@opensolaris.org http://mail.opensolaris.org/mailman/listinfo/zfs-discuss
Re: [zfs-discuss] cannot replace c10t0d0 with c10t0d0: device is too small
On Mar 4, 2011, at 11:19 AM, Eric D. Mudama wrote: On Fri, Mar 4 at 9:22, Robert Hartzell wrote: In 2007 I bought 6 WD1600JS 160GB sata disks and used 4 to create a raidz storage pool and then shelved the other two for spares. One of the disks failed last night so I shut down the server and replaced it with a spare. When I tried to zpool replace the disk I get: zpool replace tank c10t0d0 cannot replace c10t0d0 with c10t0d0: device is too small The 4 original disk partition tables look like this: Current partition table (original): Total disk sectors available: 312560317 + 16384 (reserved sectors) Part TagFlag First Sector Size Last Sector 0usrwm34 149.04GB 312560350 1 unassignedwm 0 0 0 2 unassignedwm 0 0 0 3 unassignedwm 0 0 0 4 unassignedwm 0 0 0 5 unassignedwm 0 0 0 6 unassignedwm 0 0 0 8 reservedwm 3125603518.00MB 312576734 Spare disk partition table looks like this: Current partition table (original): Total disk sectors available: 312483549 + 16384 (reserved sectors) Part TagFlag First Sector Size Last Sector 0usrwm34 149.00GB 312483582 1 unassignedwm 0 0 0 2 unassignedwm 0 0 0 3 unassignedwm 0 0 0 4 unassignedwm 0 0 0 5 unassignedwm 0 0 0 6 unassignedwm 0 0 0 8 reservedwm 3124835838.00MB 312499966 So it seems that two of the disks are slightly different models and are about 40mb smaller then the original disks. One comment: The IDEMA LBA01 spec size of a 160GB device is 312,581,808 sectors. Instead of those WD models, where neither the old nor new drives follow the IDEMA recommendation, consider buying a drive that reports that many sectors. Almost all models these days should be following the IDEMA recommendations due to all the troubles people have had. --eric -- Eric D. Mudama edmud...@bounceswoosh.org Thats encouraging, if I have to I would rather buy one new disk then 4. Thanks, Robert -- Robert Hartzell b...@rwhartzell.net RwHartzell.Net, Inc. ___ zfs-discuss mailing list zfs-discuss@opensolaris.org http://mail.opensolaris.org/mailman/listinfo/zfs-discuss
Re: [zfs-discuss] How do I Import rpool to an alternate location?
On 08/16/10 10:38 PM, George Wilson wrote: Robert Hartzell wrote: On 08/16/10 07:47 PM, George Wilson wrote: The root filesystem on the root pool is set to 'canmount=noauto' so you need to manually mount it first using 'zfs mount dataset name'. Then run 'zfs mount -a'. - George mounting the dataset failed because the /mnt dir was not empty and zfs mount -a failed I guess because the first command failed. It's possible that as part of the initial import that one of the mount points tried to create a directory under /mnt. You should first unmount everything associated with that pool, then ensure that /mnt is empty and mount the root filesystem first. Don't mount anything else until the root is mounted. - George Awesome! That worked... just recovered 100GB of data! Thanks for the help -- Robert Hartzell b...@rwhartzell.net RwHartzell.Net ___ zfs-discuss mailing list zfs-discuss@opensolaris.org http://mail.opensolaris.org/mailman/listinfo/zfs-discuss
[zfs-discuss] How do I Import rpool to an alternate location?
I have a disk which is 1/2 of a boot disk mirror from a failed system that I would like to extract some data from. So i install the disk to a test system and do: zpool import -R /mnt -f rpool bertha which gives me: bertha102G 126G84K /mnt/bertha bertha/ROOT 34.3G 126G19K legacy bertha/ROOT/snv_134 34.3G 126G 10.9G /mnt bertha/Vbox 46.9G 126G 46.9G /mnt/export/Vbox bertha/dump 2.00G 126G 2.00G - bertha/export8.05G 126G31K /mnt/export bertha/export/home 8.05G 52.0G 8.01G /mnt/export/home bertha/mail 1.54M 5.00G 1.16M /mnt/var/mail bertha/swap 4G 130G 181M - bertha/zones 6.86G 126G24K /mnt/export/zones bertha/zones/bz1 6.05G 126G24K /mnt/export/zones/bz1 bertha/zones/bz1/ROOT6.05G 126G21K legacy bertha/zones/bz1/ROOT/zbe6.05G 126G 6.05G legacy bertha/zones/bz2 821M 126G24K /mnt/export/zones/bz2 bertha/zones/bz2/ROOT 821M 126G21K legacy bertha/zones/bz2/ROOT/zbe 821M 126G 821M legacy cd /mnt ; ls bertha export var ls bertha boot etc where is the rest of the file systems and data? -- Robert Hartzell b...@rwhartzell.net RwHartzell.Net ___ zfs-discuss mailing list zfs-discuss@opensolaris.org http://mail.opensolaris.org/mailman/listinfo/zfs-discuss
Re: [zfs-discuss] How do I Import rpool to an alternate location?
On 08/16/10 07:39 PM, Mark Musante wrote: On 16 Aug 2010, at 22:30, Robert Hartzell wrote: cd /mnt ; ls bertha export var ls bertha boot etc where is the rest of the file systems and data? By default, root filesystems are not mounted. Try doing a zfs mount bertha/ROOT/snv_134 This didn't work,,, pfexec zfs mount bertha/ROOT/snv_134 cannot mount '/mnt': directory is not empty do I need set the mount point to a different location? -- Robert Hartzell b...@rwhartzell.net RwHartzell.Net ___ zfs-discuss mailing list zfs-discuss@opensolaris.org http://mail.opensolaris.org/mailman/listinfo/zfs-discuss
Re: [zfs-discuss] How do I Import rpool to an alternate location?
On 08/16/10 07:47 PM, George Wilson wrote: The root filesystem on the root pool is set to 'canmount=noauto' so you need to manually mount it first using 'zfs mount dataset name'. Then run 'zfs mount -a'. - George mounting the dataset failed because the /mnt dir was not empty and zfs mount -a failed I guess because the first command failed. -- Robert Hartzell b...@rwhartzell.net RwHartzell.Net ___ zfs-discuss mailing list zfs-discuss@opensolaris.org http://mail.opensolaris.org/mailman/listinfo/zfs-discuss