ramble
asked on
Solaris 9 - 3510 - sccli - recreating a concatenation
I haven't done much with metadb or Solstice DiskSuite (I guess now called "Solaris Volume Manager" ("SVM") )
I'm not even sure if I need to use them...
I have a StorEdge 3510 that has several disks in it. 3 disks are concatentated together as one drive. Well, one of the drives died in that concatenation, so I replaced the drive. Ofcourse, because there wasn't any fault tolerance, I lost all the data. So, what I want to do is simply recreate the concatenation with the 3 drives.
So, I need:
1. determine which 3 drives were used in the previous concat.
2. recreate this concat
3. format it - if needed, so I can dump some data to it.
I'm guessing that I don't even use the metadb or SVM, and I recreate this within the sccli tool.
Thanks for any input!
I'm not even sure if I need to use them...
I have a StorEdge 3510 that has several disks in it. 3 disks are concatentated together as one drive. Well, one of the drives died in that concatenation, so I replaced the drive. Ofcourse, because there wasn't any fault tolerance, I lost all the data. So, what I want to do is simply recreate the concatenation with the 3 drives.
So, I need:
1. determine which 3 drives were used in the previous concat.
2. recreate this concat
3. format it - if needed, so I can dump some data to it.
I'm guessing that I don't even use the metadb or SVM, and I recreate this within the sccli tool.
Thanks for any input!
ASKER
The sccli tool is the 3510 StorEdge Array manager...
if you use disk suite (or SVM) you should see a failed component:
# metastat
or
# metastat <mdev>
and a message explaining what to do to fix the problem (usually a "metareplace")
If you see something like "metareplace <mdev> <comp1> <comp2>" you may want to run
# metareplace -e <mdev> <comp1>
to have it use the same component as before (as you did replace it)
# metastat
or
# metastat <mdev>
and a message explaining what to do to fix the problem (usually a "metareplace")
If you see something like "metareplace <mdev> <comp1> <comp2>" you may want to run
# metareplace -e <mdev> <comp1>
to have it use the same component as before (as you did replace it)
ASKER
The sccli tool is the 3510 StorEdge Array manager...I believe.
So, I thought that the sccli tool would HAVE to be used to manage the disks inside the array.
Is that not the case?
So, I thought that the sccli tool would HAVE to be used to manage the disks inside the array.
Is that not the case?
I doubt you use SDS (or LVM) as the 3510 has hardware RAID
Using sccli oyu may want to check your logical drive settings with
# sccli <device> show logical-drives
Everything else is left to using "sccli" to reconfigure the RAID
Cheers
Using sccli oyu may want to check your logical drive settings with
# sccli <device> show logical-drives
Everything else is left to using "sccli" to reconfigure the RAID
Cheers
ASKER
Ok...using you command, I get:
sccli> show logical-drives
LD LD-ID Size Assigned Type Disks Spare Failed Status
-------------------------- ---------- ---------- ---------- ---------- -------
ld0 5F0AFF2E 204.35GB Primary RAID1 6 1 0 Good
ld1 2125D7FF 68.12GB Primary RAID1 2 1 0 Good
ld2 1A3AC015 204.35GB Primary RAID0 2 0 0 Incomplete
So, it looks like it's the ld2 that needs to be rebuilt...
The concatenation was using 3 disks...3 70GB disks, making it approximately 200 GB.
do you know what I need to do from here?
sccli> show logical-drives
LD LD-ID Size Assigned Type Disks Spare Failed Status
--------------------------
ld0 5F0AFF2E 204.35GB Primary RAID1 6 1 0 Good
ld1 2125D7FF 68.12GB Primary RAID1 2 1 0 Good
ld2 1A3AC015 204.35GB Primary RAID0 2 0 0 Incomplete
So, it looks like it's the ld2 that needs to be rebuilt...
The concatenation was using 3 disks...3 70GB disks, making it approximately 200 GB.
do you know what I need to do from here?
ASKER
...and, from the looks at this command, it looks like 8,9,10 are the ones I need to concatenate.
10 looks likes its the failed disk, I replaced...breaking the 8,9,10 concatenation.
sccli> show disks
Ch Id Size Speed LD Status IDs
-------------------------- ---------- ---------- ---------- ---------- -----
2 0 68.37GB 200MB ld0 ONLINE SEAGATE ST373307FSUN72G 0307
S/N 5GZ5D47G01108327
2 1 68.37GB 200MB ld0 ONLINE SEAGATE ST373307FSUN72G 0307
S/N 5GZ5CTY801108327
2 2 68.37GB 200MB ld0 ONLINE SEAGATE ST373307FSUN72G 0307
S/N 5GZ5D3Q701108327
2 3 68.37GB 200MB ld0 ONLINE SEAGATE ST373307FSUN72G 0307
S/N 5GZ5AA0401108325
2 4 68.37GB 200MB ld0 ONLINE SEAGATE ST373307FSUN72G 0307
S/N 5GZ5AAZE01108326
2 5 68.37GB 200MB ld0 ONLINE SEAGATE ST373307FSUN72G 0307
S/N 5GZ59E1X01108325
2 6 68.37GB 200MB ld1 ONLINE SEAGATE ST373307FSUN72G 0307
S/N 5GZ5ADQ601108326
2 7 68.37GB 200MB ld1 ONLINE SEAGATE ST373307FSUN72G 0307
S/N 5GZ5CXS301108327
2 8 68.37GB 200MB ld2 ONLINE SEAGATE ST373307FSUN72G 0307
S/N 5GZ5D4AQ01108327
2 9 68.37GB 200MB ld2 ONLINE SEAGATE ST373307FSUN72G 0307
S/N 5GZ5AA5C01108326
2 10 68.37GB 200MB NONE FRMT SEAGATE ST373307FSUN72G 0307
S/N 5GZ5LS7101108320
2 11 68.37GB 200MB GLOBAL STAND-BY SEAGATE ST373307FSUN72G 0307
S/N 5GZ5ABDA01108326
10 looks likes its the failed disk, I replaced...breaking the 8,9,10 concatenation.
sccli> show disks
Ch Id Size Speed LD Status IDs
--------------------------
2 0 68.37GB 200MB ld0 ONLINE SEAGATE ST373307FSUN72G 0307
S/N 5GZ5D47G01108327
2 1 68.37GB 200MB ld0 ONLINE SEAGATE ST373307FSUN72G 0307
S/N 5GZ5CTY801108327
2 2 68.37GB 200MB ld0 ONLINE SEAGATE ST373307FSUN72G 0307
S/N 5GZ5D3Q701108327
2 3 68.37GB 200MB ld0 ONLINE SEAGATE ST373307FSUN72G 0307
S/N 5GZ5AA0401108325
2 4 68.37GB 200MB ld0 ONLINE SEAGATE ST373307FSUN72G 0307
S/N 5GZ5AAZE01108326
2 5 68.37GB 200MB ld0 ONLINE SEAGATE ST373307FSUN72G 0307
S/N 5GZ59E1X01108325
2 6 68.37GB 200MB ld1 ONLINE SEAGATE ST373307FSUN72G 0307
S/N 5GZ5ADQ601108326
2 7 68.37GB 200MB ld1 ONLINE SEAGATE ST373307FSUN72G 0307
S/N 5GZ5CXS301108327
2 8 68.37GB 200MB ld2 ONLINE SEAGATE ST373307FSUN72G 0307
S/N 5GZ5D4AQ01108327
2 9 68.37GB 200MB ld2 ONLINE SEAGATE ST373307FSUN72G 0307
S/N 5GZ5AA5C01108326
2 10 68.37GB 200MB NONE FRMT SEAGATE ST373307FSUN72G 0307
S/N 5GZ5LS7101108320
2 11 68.37GB 200MB GLOBAL STAND-BY SEAGATE ST373307FSUN72G 0307
S/N 5GZ5ABDA01108326
ASKER CERTIFIED SOLUTION
membership
This solution is only available to members.
To access this solution, you must be a member of Experts Exchange.
ASKER
JustUNIX:
From the link you posted above, (under Resolution:)
I'm suppose to select "view and edit SCSI Drives" from the menu...
- go "view and edit SCSI Drives"
How do I get to the menu? There seems to be a program that brings up some kind of "menu"...what program (or executable) is it?
From the link you posted above, (under Resolution:)
I'm suppose to select "view and edit SCSI Drives" from the menu...
- go "view and edit SCSI Drives"
How do I get to the menu? There seems to be a program that brings up some kind of "menu"...what program (or executable) is it?
check the tools that come with your RAID system. Try something like:
grep `sccli /var/sadm/install/contents | awk '{print $4]'` /var/sadm/install/contents
The meu is available in the 3510 itself (log on to the RAID box itself - is it connected to the network?)
Cheers
grep `sccli /var/sadm/install/contents
The meu is available in the 3510 itself (log on to the RAID box itself - is it connected to the network?)
Cheers
ASKER
http://www.sun.com/products-n-solutions/hardware/docs/html/817-3711-10/preface.html
"chapter 2" seems to suggest connecting to the array through a COM port.
I'm guess that, that will get me to that "menu" screen that I need to see.
#1 - will I have to bring the array completely down - just to connect with a terminal progam? Or, can I connect to it 'hot'?
#2 - Is this basically the only way to get to the array 'menu'? It sounds awefully inconvenient - other arrays I've configured allow you to manage it on the fly.
I see that I could also manage it through it's own IP address. I guess I could connect it to the net - but this brings me question, how do I assign an IP address to it - and do I use telnet to access it?
"chapter 2" seems to suggest connecting to the array through a COM port.
I'm guess that, that will get me to that "menu" screen that I need to see.
#1 - will I have to bring the array completely down - just to connect with a terminal progam? Or, can I connect to it 'hot'?
#2 - Is this basically the only way to get to the array 'menu'? It sounds awefully inconvenient - other arrays I've configured allow you to manage it on the fly.
I see that I could also manage it through it's own IP address. I guess I could connect it to the net - but this brings me question, how do I assign an IP address to it - and do I use telnet to access it?
ASKER
From my original question:
1. determine which 3 drives were used in the previous concat.
2. recreate this concat
3. format it - if needed, so I can dump some data to it.
Here are the steps I've taken to recreate the concatenation:
-=-=-=-=-=-=-=-=-=-=-=-=-= -=-=-=-=-= -=-=-=-=-= -=-=-=-=-=
1. determine which 3 drives were used in the previous concat.
Connect to the 3510 array manager program:
The 3510 array has a COM port that must be connected to the server.
To connect to it via COM1, type:
tip -38400 /dev/ttya
To connect to it via COM2, type:
tip -38400 /dev/ttyb
Once connected, ctrl-l to refresh the display.
In the "view and edit scsi Drives", I could see the specific drive that was replaced. The "status" column wasn't ON-LINE like the other drives.
In the "view and edit Logical drives", the "status" of the logical set was INCOMPLETE. When selecting this set, I could see the exact 3 drives involved in making up the logical drive/concatenation.
2. recreate this concat
view and edit Host Luns: deleted the drives associated with the affected lun.
view and edit logical Volumes: deleted the Volume associated with the affected volume.
view and edit logical drives: deleted the logical drive associated with the affected drive.
Recreate the concatenation is reverse order:
view and edit logical drives
view and edit logical volumes
view and edit host luns...
3. format it - if needed, so I can dump some data to it.
newfs /dev/dsk/cXXXXXXX
where XXXXXXX is the specific controller/disk to prepare.
-=-=-=-=-=-=-=-=-=-=-=-=-= -=-=-=-=-= -=-=-=-=-= -=-=-=-=-= -=
Some things to note:
You may have to power off the array, after replacing the disk. Then do a "scan" on it - to be recognized by the array.
1. determine which 3 drives were used in the previous concat.
2. recreate this concat
3. format it - if needed, so I can dump some data to it.
Here are the steps I've taken to recreate the concatenation:
-=-=-=-=-=-=-=-=-=-=-=-=-=
1. determine which 3 drives were used in the previous concat.
Connect to the 3510 array manager program:
The 3510 array has a COM port that must be connected to the server.
To connect to it via COM1, type:
tip -38400 /dev/ttya
To connect to it via COM2, type:
tip -38400 /dev/ttyb
Once connected, ctrl-l to refresh the display.
In the "view and edit scsi Drives", I could see the specific drive that was replaced. The "status" column wasn't ON-LINE like the other drives.
In the "view and edit Logical drives", the "status" of the logical set was INCOMPLETE. When selecting this set, I could see the exact 3 drives involved in making up the logical drive/concatenation.
2. recreate this concat
view and edit Host Luns: deleted the drives associated with the affected lun.
view and edit logical Volumes: deleted the Volume associated with the affected volume.
view and edit logical drives: deleted the logical drive associated with the affected drive.
Recreate the concatenation is reverse order:
view and edit logical drives
view and edit logical volumes
view and edit host luns...
3. format it - if needed, so I can dump some data to it.
newfs /dev/dsk/cXXXXXXX
where XXXXXXX is the specific controller/disk to prepare.
-=-=-=-=-=-=-=-=-=-=-=-=-=
Some things to note:
You may have to power off the array, after replacing the disk. Then do a "scan" on it - to be recognized by the array.
use metastat to find out what the components were
i.e.
metastat d10
then
metaclear d10
then
metainit d10 1 3 c1t0d0s2 c1t1d0s2 c1t2d0s2
(or whatever your three slices are)
after this
newfs /dev/md/rdsk/d10
and thats about it