[prev in list] [next in list] [prev in thread] [next in thread] 

List:       opensolaris-lvm-discuss
Subject:    Re: [lvm-discuss] [Q] What is causing metainit to complain with "no
From:       Maurice Volaski <mvolaski () aecom ! yu ! edu>
Date:       2008-07-23 16:11:03
Message-ID: p06240401c4ad068d8ae9 () [129 ! 98 ! 90 ! 227]
[Download RAW message or body]

>Maybe the sector count for the disk is too small ?
>
>Have a look at the libmeta source for meta_check_stripe or 
>meta_create_stripe as both can return ENOSPC.
>
>Attached is a dtrace script that might help (it prints
>a stack of the calling functions).
>
>Run as:
>
>dtrace -q -c "metainit d20 1 1 c1t21d0s1" -s meta.d
>

Thank you for your response. Here is the result:

metagetstart return: from: 0xe8 start_blk: 0

               libmeta.so.1`metagetstart+0xe8
               libmeta.so.1`meta_create_stripe+0x5f
               libmeta.so.1`meta_init_stripe+0x3fe
               libmeta.so.1`meta_init_name+0x16f
               metainit`init_name+0xfd
               metainit`main+0x2a8
               metainit`_start+0x7a
metagetsize return: from: 0x72 size: 128

               libmeta.so.1`metagetsize+0x72
               libmeta.so.1`meta_create_stripe+0x5f
               libmeta.so.1`meta_init_stripe+0x3fe
               libmeta.so.1`meta_init_name+0x16f
               metainit`init_name+0xfd
               metainit`main+0x2a8
               metainit`_start+0x7a
meta_check_stripe return: from: 0x202 value: 4294967295

               libmeta.so.1`meta_check_stripe+0x202
               libmeta.so.1`meta_init_stripe+0x3fe
               libmeta.so.1`meta_init_name+0x16f
               metainit`init_name+0xfd
               metainit`main+0x2a8
               metainit`_start+0x7a
Result: size(0 - 0): 0

I'm not exactly sure, but it appears that at first, metainit sees the 
correct slice and reports it as size 128, but in the last line above, 
it seems to think it is zero.

I'm not sure of the significance the check stripe value.

Finally, if it did see the size as 128, it's not clear that it is 
smaller than some minimum.

>
>
>Maurice Volaski wrote:
>>I am trying to setup a mirror of slices that will be used to house 
>>the bitmap volumes for AVS, but metainit is failing with:
>>
>>metainit: nevada1: c1t1d0s1: No space left on device
>>
>>The following volumes will be part of AVS and they each have two 
>>formatted slices, one where ZFS will go and the other where the 
>>bitmap volume for AVS will go.
>>c1t1d0
>>c1t2d0
>>c1t3d0
>>c1t4d0
>>
>>The instructions at 
>>http://blogs.sun.com/AVS/entry/avs_and_zfs_seamless suggest 
>>mirroring the bitmap volumes.
>>
>>So I created a set of database replicas	flags		first 
>>blk	block count
>>      a        u    	16		8192		/dev/dsk/c1t5d0s0
>>      a        u    	8208		8192		/dev/dsk/c1t5d0s0
>>      a        u    	16		8192		/dev/dsk/c1t5d0s1
>>      a        u    	8208		8192		/dev/dsk/c1t5d0s1
>>
>>Now here is how c1t1d0 looks
>>*                          First     Sector    Last
>>* Partition  Tag  Flags    Sector     Count    Sector  Mount Directory
>>        0      4    00         34  12566111  12566144
>>        1      4    00   12566145       128  12566272
>>        8     11    00   12566495     16384  12582878
>>
>>and c1t3d0
>>*                          First     Sector    Last
>>* Partition  Tag  Flags    Sector     Count    Sector  Mount Directory
>>        0      4    00         34  12566111  12566144
>>        1      4    00   12566145       128  12566272
>>        8     11    00   12566495     16384  12582878
>>
>>The slice 0's will be used by ZFS and slice 1's will be used by AVS.
>>
>>The instructions suggest first putting the bitmap volume slice of 
>>the odd numbered volumes, t1 and t3, into a submirror and then 
>>doing the same for the even numbered ones.
>>
>>But when I manually try it for the odd-numbered volumes with
>>metainit d1 2 1 c1t1d0s1 1 c1t3d0s1, I get:
>>metainit: nevada1: c1t1d0s1: No space left on device

-- 

Maurice Volaski, mvolaski@aecom.yu.edu
Computing Support, Rose F. Kennedy Center
Albert Einstein College of Medicine of Yeshiva University
_______________________________________________
lvm-discuss mailing list
lvm-discuss@opensolaris.org
[prev in list] [next in list] [prev in thread] [next in thread]