Should I be Worried: sda: p1 size .... limited to end of disk

After the fun and games with my other LSI MegaRAID card, I'm a glutton for punishment.

I'm building another system, which this time, has an LSI MegaRAID SATA 300-8X controller in it.  This appears to be working perfectly, and a lot faster than my other card, with the exception of this warning message during boot:

sda: p1 size 2930304132 limited to end of disk

Is this something I should be worried about, and how does it come about.  I built a new RAID 5, Logical device, on the card, which the LSI utility says is 1430511MB in size.

Here's the relevant output from DMESG:

scsi 4:1:0:0: Direct-Access     MegaRAID LD 0 RAID5 1430G 815C PQ: 0 ANSI: 2 sd 4:1:0:0: [sda] 2929686528 512-byte hardware sectors: (1.49 TB/1.36 TiB) sd 4:1:0:0: [sda] Write Protect is off
sd 4:1:0:0: [sda] Mode Sense: 00 00 00 00
sd 4:1:0:0: [sda] Asking for cache data failed
sd 4:1:0:0: [sda] Assuming drive cache: write through
sd 4:1:0:0: [sda] 2929686528 512-byte hardware sectors: (1.49 TB/1.36 TiB) sd 4:1:0:0: [sda] Write Protect is off
sd 4:1:0:0: [sda] Mode Sense: 00 00 00 00
sd 4:1:0:0: [sda] Asking for cache data failed
sd 4:1:0:0: [sda] Assuming drive cache: write through
sda: sda1
sda: p1 size 2930304132 limited to end of disk
sd 4:1:0:0: [sda] Attached SCSI disk
sd 4:1:0:0: Attached scsi generic sg1 type 0

And /proc/partions:

major minor  #blocks  name

 3        0   20044080 hda
 3        1    9727326 hda1
 3        2    9727357 hda2
 3        4     586372 hda4
 8        0 1464843264 sda
 8        1 1464843232 sda1
253        0  314572800 dm-0
253        1  157286400 dm-1
253        2  209715200 dm-2
253        3  104857600 dm-3
253        4  104857600 dm-4
253        5  262144000 dm-5

And fdisk:

Disk /dev/sda: 1499.9 GB, 1499999502336 bytes
255 heads, 63 sectors/track, 182364 cylinders
Units = cylinders of 16065 * 512 = 8225280 bytes
Disk identifier: 0x0003d809

 Device Boot      Start         End      Blocks   Id  System /dev/sda1               1      182403  1465152066   8e  Linux LVM

So, where is the discrepancy creeping in.

Also, I see this in DMESG:

sd 4:1:0:0: [sda] Assuming drive cache: write through

But, as this card has a battery backup, I have it set to "Write Back" in the configuration.  So, what mode is it really operating in.

Cheers,
Eddie

--
To unsubscribe from this list: send the line "unsubscribe linux-scsi" in the body of a message to [email protected]
More majordomo info at  http://vger.kernel.org/majordomo-info.html

Quoted 10 lines · Eddie

After the fun and games with my other LSI MegaRAID card, I'm a glutton for punishment.

I'm building another system, which this time, has an LSI MegaRAID SATA 300-8X controller in it. This appears to be working perfectly, and a lot faster than my other card, with the exception of this warning message during boot:

sda: p1 size 2930304132 limited to end of disk

sda only has 2929686528 blocks - "p1" is somehow defined to go past the end of the RAID "device".

I've no clue where the "2930304132" is coming from. I'd want to see the "raw" values the partition table.
Seems like the partition table was written with the wrong size,

Quoted 4 lines

Is this something I should be worried about, and how does it come about. I built a new RAID 5, Logical device, on the card, which the LSI utility says is 1430511MB in size.

1430511*1024*2
2929686528

Quoted 43 lines

Here's the relevant output from DMESG:

scsi 4:1:0:0: Direct-Access MegaRAID LD 0 RAID5 1430G 815C PQ: 0 ANSI: 2
sd 4:1:0:0: [sda] 2929686528 512-byte hardware sectors: (1.49 TB/1.36 TiB) sd 4:1:0:0: [sda] Write Protect is off
sd 4:1:0:0: [sda] Mode Sense: 00 00 00 00
sd 4:1:0:0: [sda] Asking for cache data failed
sd 4:1:0:0: [sda] Assuming drive cache: write through
sd 4:1:0:0: [sda] 2929686528 512-byte hardware sectors: (1.49 TB/1.36 TiB) sd 4:1:0:0: [sda] Write Protect is off
sd 4:1:0:0: [sda] Mode Sense: 00 00 00 00
sd 4:1:0:0: [sda] Asking for cache data failed
sd 4:1:0:0: [sda] Assuming drive cache: write through
sda: sda1
sda: p1 size 2930304132 limited to end of disk
sd 4:1:0:0: [sda] Attached SCSI disk
sd 4:1:0:0: Attached scsi generic sg1 type 0

And /proc/partions:

major minor #blocks name

 3        0   20044080 hda
 3        1    9727326 hda1
 3        2    9727357 hda2
 3        4     586372 hda4
 8        0 1464843264 sda
 8        1 1464843232 sda1
253        0  314572800 dm-0
253        1  157286400 dm-1
253        2  209715200 dm-2
253        3  104857600 dm-3
253        4  104857600 dm-4
253        5  262144000 dm-5

And fdisk:

Disk /dev/sda: 1499.9 GB, 1499999502336 bytes
255 heads, 63 sectors/track, 182364 cylinders
Units = cylinders of 16065 * 512 = 8225280 bytes
Disk identifier: 0x0003d809

Quoted 5 lines

Device Boot Start End Blocks Id System
/dev/sda1 1 182403 1465152066 8e Linux LVM

So, where is the discrepancy creeping in.

Can you print this using "Units of Blocks" instead of "Cyclinders"? Use "-u" parameter.

hth,
grant

Quoted 12 lines

Also, I see this in DMESG:

sd 4:1:0:0: [sda] Assuming drive cache: write through

But, as this card has a battery backup, I have it set to "Write Back" in the configuration. So, what mode is it really operating in.

Cheers,
Eddie

The raw partition table entry is:

00 01 01 00 8E FE FF FF 3F 00 00 00 84 E4 A8 AE

Active Flag = 00
Starting Head = 01
Starting Cylinder = 0001
Partition Type = x'8E' - Linux LVM
Ending Head = x'FE' - 254
Ending Cylinder/Sector = x'FFFF' - LBA
Offset to Start = x'0000003F' - 63

# of Sectors = x'AEA8E484' - 2930304132 - Hmmmmmmm

So, that looks like the culprit. Now, how did it happen. I know I can, sort of safely, rebuild that, by running (c)fdisk and making sure I start the partition at exactly the same point. But all I did the first time, was use cfdisk, and allocate the complete disk to the 1st primary partition.

 Device Boot      Start         End      Blocks   Id  System /dev/sda1              63  2929677659  1464838798+  8e  Linux LVM

OK, the total sectors seems to match the hardware scan -> sd 4:1:0:0: [sda] 2929686528 512-byte hardware sectors: (1.49 TB/1.36 TiB)

But 255 * 63 * 182364 = 2929677660.  What's this discrepancy due to.  Does the drive report ALL sectors, even the ones used to replace "bad" ones.

However, 2929677660 does seem to tie up quite nicely with the "end sector", from the fdisk, so I'm assuming that everything is now correct.

Well, apart from the fact that LVM reports a Total PE allocation of 357627 extents, which maps to 1464840192 Blocks, according to my maths anyway, which happens to be bigger than the "available" sectors on the disk.  I guess it's time to hit the LVM list, to see if there's any way I can safely correct this, before anything tries to use the "phantom" sectors.

Cheers,
Eddie