Sponsored Content
Operating Systems Solaris metadevice submirrors 0 and 1 not consistent Post 302431600 by Max Power on Tuesday 22nd of June 2010 08:38:58 AM
Old 06-22-2010
metadevice submirrors 0 and 1 not consistent

At a new job, I've inherited a handful of Solaris 10 SunFire V240 servers.

I noticed that the metadevices are not consistent, in a way that I've not seen before....

internal mirrored disks:

metadevice d60 has c1t0d0s6 as submirror 0 (d61) and c1t1d0s6 as submirror 1 (d62)

metadevice d40 has c1t1d0s4 as submirror 0 (d42) and c1t0d0s4 as submirror 1 (d41)

etc....

Is this a problem? It is certainly strange. Half the mirrors have disk c1t0d0 as submirror 0 and c1t1d0 as submirror 1. The other half of the mirrors have disk c1t1d0 as submirror 0 and disk c1t0d0 as submirror 1.

I think I know how it happened:
When the previous sysadmin installed recommended patches, the mirrors were broken (some, but not all) and then a few days later the devices were synced (metattach). By only breaking some of the mirrors to install the patches, and then rebooting, things got "reversed" in metadeviceland.

Is this a problem? What happens if disk c0t0d0 dies, and needs to be replaced?


Here is an example. 3 metadevices on this system are one way, and 3 are the other way, with respect to their submirrors:


Code:
$ metastat | more
d60: Mirror
    Submirror 0: d61
      State: Okay
    Submirror 1: d62
      State: Okay
    Pass: 1
    Read option: roundrobin (default)
    Write option: parallel (default)
    Size: 8395200 blocks (4.0 GB)

d61: Submirror of d60
    State: Okay
    Size: 8395200 blocks (4.0 GB)
    Stripe 0:
        Device     Start Block  Dbase        State Reloc Hot Spare
        c1t0d0s6          0     No            Okay   Yes

d62: Submirror of d60
    State: Okay
    Size: 8395200 blocks (4.0 GB)
    Stripe 0:
        Device     Start Block  Dbase        State Reloc Hot Spare
        c1t1d0s6          0     No            Okay   Yes

d40: Mirror
    Submirror 0: d42
      State: Okay
    Submirror 1: d41
      State: Okay
    Pass: 1
    Read option: roundrobin (default)
    Write option: parallel (default)
    Size: 16780224 blocks (8.0 GB)

d42: Submirror of d40
    State: Okay
    Size: 16780224 blocks (8.0 GB)
    Stripe 0:
        Device     Start Block  Dbase        State Reloc Hot Spare
        c1t1d0s4          0     No            Okay   Yes

d41: Submirror of d40
    State: Okay
    Size: 16780224 blocks (8.0 GB)
    Stripe 0:
        Device     Start Block  Dbase        State Reloc Hot Spare
        c1t0d0s4          0     No            Okay   Yes


Last edited by Scott; 06-22-2010 at 10:11 AM.. Reason: Please use code tags
 

9 More Discussions You Might Find Interesting

1. UNIX for Advanced & Expert Users

keep UIDs/GIDs consistent

Hi, What is the best ways to keep UIDs and GIDs consistent across unix and linux server. my company have a servers running on hpux, linux, aix and many of them have veritas cluster and hacmp running, many time user account have been created only on one of the cluster node and not the others... (4 Replies)
Discussion started by: robertngo
4 Replies

2. Shell Programming and Scripting

Trouble getting consistent PID of tomcat server from ps

Hello all, I am working on a script to automate the process of restarting some tomcat servers. The versions of tomcat are pretty old and it often happens that they don't shut down gracefully and the PID needs to be manually killed. I am having a bear of a time getting a consistent PID from ps... (1 Reply)
Discussion started by: dkaplowitz
1 Replies

3. Shell Programming and Scripting

Negate alerting for particular metadevice.

Hi All... I have a script that checks for any problems(particularly looks for 'Needs Maintenance') with metadevices and alerts accordingly. This was not configured to alert for a particular metadevice. Now i want to negate alerting for a particular metadevice(say d40). Is this possible? I am... (6 Replies)
Discussion started by: reddybs
6 Replies

4. Solaris

Unable to mount Metadevice

Hi, I am running Solaris 8 on a V880 which has recently crashed. It has got a fibre connection to a StorageTrek array containing my database info. Since the crash I have been unable to mount the 2 metadevices which correspond to /spool & /apps - these FS's contain my db info and theredore I... (20 Replies)
Discussion started by: Duffs
20 Replies

5. Solaris

metadevice problem

hi all i am very new to solaris...i am trying configure raid 0 in my ultrascsi box which is having 6 sun36GB hard disks. I have created 4 state database replicas in the following slices /dev/dsk/c1t1d0s7 /dev/dsk/c1t2d0s7 /dev/dsk/c1t3d0s7 and /dev/dsk/c1t4d0s7. i could not use any other... (7 Replies)
Discussion started by: kingston
7 Replies

6. Shell Programming and Scripting

How to echo columns with consistent spacing?

for (.....) echo -e "$Name | $Age | $Sex\t|$Grade\t\t" done output: John |12 |Male |6th Jack |15 |Male |8th Zachary |15 |Male |9th I want the lines to line up...but it's out of line when the name is long (2 Replies)
Discussion started by: etranman1
2 Replies

7. Solaris

Submirrors swapped after metattach

We had few mirror disks e.g d0 had two submirrors Submirror 0: d2 Submirror 1: d1 I opened mirrors for some activity. by command metadetach d0 d2 After activity I reattach mirrors with command metattach d0 d2 But after that strangely sumirror position is swaped ... (3 Replies)
Discussion started by: kashif_islam
3 Replies

8. Solaris

Metadevice Too Small To Attach

Hi All, I am having trouble with my metadevices and I hope someone can help. when i try to attach my submirror i get the following error: root@xxxx>metattach d31 d21 metattach: tbair1: d21: submirror too small to attach root@xxxx> However, There is a submirror on the other disk which is... (3 Replies)
Discussion started by: A-Train
3 Replies

9. UNIX for Beginners Questions & Answers

Scripts dont give a consistent output..

I have at moment created a setup script that is capable of setting up a workspace for an kaldi enviroment. The script can be found here setup_base_file I guess you wouldn't not be able to run it, without having Kaldi installed, but since this question more relates to scripting, than the... (9 Replies)
Discussion started by: kidi
9 Replies
metasync(1M)						  System Administration Commands					      metasync(1M)

NAME
metasync - handle metadevice resync during reboot SYNOPSIS
/usr/sbin/metasync -h /usr/sbin/metasync [-s setname] [buffer_size] metadevice /usr/sbin/metasync [-s setname] -r [buffer_size] /usr/sbin/metasync -p metadevice DESCRIPTION
The metasync command starts a resync operation on the specified metadevice. All components that need to be resynced are resynced. If the system crashes during a RAID5 initialization, or during a RAID5 resync, either an initialization or resync restarts when the system reboots. Applications are free to access a metadevice at the same time that it is being resynced by metasync. Also, metasync performs the copy oper- ations from inside the kernel, which makes the utility more efficient. Use the -r option in boot scripts to resync all possible submirrors. OPTIONS
The following options are supported: -h Displays usage message. -p metadevice Regenerates parity information for RAID5 metadevices. -s setname Specifies the name of the diskset on which metasync will work. Using the -s option will cause the command to perform its administrative function within the specified diskset. Without this option, the command will perform its function on local metadevices. -r Specifies that the metasync command handle special resync requirements during a system reboot. metasync -r should only be invoked from the svc:/system/mdmonitor service. The metasync command only resyncs those metadevices that need to be resynced. metasync schedules all the mirror resyncs according to their pass numbers. To override the default buffer_size value used by the svc:/system/mdmonitor service, you can edit /etc/system to specify: set md_mirror:md_resync_bufsz = 2048 so that resyncs occur as quickly as possible. OPERANDS
buffer_size Specifies the size (number of 512-byte disk blocks) of the internal copy buffer for the mirror resync. The size defaults to 128 512-byte disk blocks (64 Kbytes). It can be no more than 2048 blocks. For best performance (quick- est completion of the resync), 2048 blocks is the recommended size. EXIT STATUS
The following exit values are returned: 0 Successful completion. >0 An error occurred. ATTRIBUTES
See attributes(5) for descriptions of the following attributes: +-----------------------------+-----------------------------+ | ATTRIBUTE TYPE | ATTRIBUTE VALUE | +-----------------------------+-----------------------------+ |Availability |SUNWmdu | +-----------------------------+-----------------------------+ SEE ALSO
mdmonitord(1M), metaclear(1M), metadb(1M), metadetach(1M), metahs(1M), metainit(1M), metaoffline(1M), metaonline(1M), metaparam(1M), metarecover(1M), metarename(1M), metareplace(1M), metaroot(1M), metaset(1M), metassist(1M), metastat(1M), metattach(1M), md.tab(4), md.cf(4), mddb.cf(4), md.tab(4), attributes(5), md(7D) Solaris Volume Manager Administration Guide NOTES
The metasync service is managed by the service management facility, smf(5), under the service identifier: svc:/system/mdmonitor Administrative actions on this service, such as enabling, disabling, or requesting restart, can be performed using svcadm(1M). The ser- vice's status can be queried using the svcs(1) command. SunOS 5.10 24 Nov 2004 metasync(1M)
All times are GMT -4. The time now is 11:03 PM.
Unix & Linux Forums Content Copyright 1993-2022. All Rights Reserved.
Privacy Policy