Disk Arrays
cancel
Showing results for 
Search instead for 
Did you mean: 

Procedure to add disk to VA7410

SOLVED
Go to solution
Patil P
Occasional Contributor

Procedure to add disk to VA7410

Hello,

We are increasing the disk/size of VA7410 and adding 4 disks in it on 8/11/05.

I want to know the step-by-step procedure of adding the disk in VA using SAM

1. Whether the disk will be formatted automatically?
2. What is a procedure to allocate space to RG1 & RG2?
3. Procedure to create LUN.
4. Procedure to increase size of present LUN.
5. What precaution to take for protecting old data.
6. Any more info

Waiting for replay.

Regards

Patil P
8 REPLIES
Ranjith_5
Honored Contributor

Re: Procedure to add disk to VA7410

Hi Patil,

At my place we had added disk. If auto include and auto format is enabled in VA added disks will be included in the disk group and will be formatted automatically.

Another thing in VA, disks included in the redundancy group is managed by the device itself. There is no manual intervention required. You can only manage the space allocation from the redundancy groups.

http://forums1.itrc.hp.com/service/forums/questionanswer.do?threadId=47827

The following is about the addition of a new disk enclosure. This will also help you.

http://h20000.www2.hp.com/bizsupport/TechSupport/Document.jsp?lang=en&cc=us&taskId=115&prodSeriesId=89018&prodTypeId=12169&objectID=lpg35228

Regards,
Syam
Torsten.
Acclaimed Contributor

Re: Procedure to add disk to VA7410

The array is not managed by SAM, you have to use commandview sdm. This software have to be installed on one connected host. You can use command line, menu based user interface or GUI.
1. get the alias name of your VA
armdsp -i

If your VA is configured to "auto include = on", than all disks you inserted will first formatted to 520 Byte blocks and included to the RG automatically. Since you *cannot* increase a LUN, you may now configure a new LUN using a user interface. Once the LUN is created, use SAM to configure LVM.

get all information:

armdsp -a

See the commandview sdm manual for more information.

Please assign some points to all answers, if you find them valueable.

Hope this helps!
Regards
Torsten.

__________________________________________________
There are only 10 types of people in the world -
those who understand binary, and those who don't.

__________________________________________________
No support by private messages. Please ask the forum!

If you feel this was helpful please click the KUDOS! thumb below!   
Ranjith_5
Honored Contributor

Re: Procedure to add disk to VA7410

Hi Patil,

Attached procedure to create LUN. Prepared specially for you just now.

See the word doc with screen shots added for easy understanding.

Regards,
Syam
Don Mallory
Trusted Contributor

Re: Procedure to add disk to VA7410

Hi Patil,

I would say that others have covered question 1.

2. Allocating space between RG1 and RG2 is pretty simple. Odd numbered disks are part of RG1, even numbered disks are part of RG2.

So, if you add all the disks to odd slots, they will all add to RG1.

Keep in mind, that running unbalanced is not recommended becuase it also unbalances the I/O over the disk channels and puts load on the fiber loops.

3. Syam's procedure for using the CommandView SDM GUI (launcher) interface is pretty sound, however there's another way that doesn't require eXceed or another X session:

Log into the node you have CommandView installed on.

Run 'armdsp -a arrayname'

If this doesn't work, or you don't know your array name, you can run, 'armdiscover', it will find the name of your array in the right hand column.

Look at the output of your armdsp -a, it will show you EVERYTHING about your array, including the LUNs. Figure out what LUN number you want to use and how much disk space you have free (check the free disk per RG near the bottom). Now create the LUN:

Run: armcfg -L LUN -a capacity -g RG arrayname

Where:

LUN = LUN number you want to create
capacity = size in KB or GB, 100G = 100GB, 1024K = 1MB
RG = Redundancy group number, 1 or 2.
arrayname = the name of your array as shown in armdiscover or used with armdsp.


Other useful arm commands:

armmgr - manage other array functions
armrbld - set rebuild options, view rebuild status
armdiag - pull diagnostic data to send to HP support.

4. Changing sizes.. the VA's aren't very good at this. you need to remove and re-create the LUN. The EVAs can make online size changes.

5. Always back up your data prior to making any major changes. The VA's are pretty stable most of the time.

6. Always use the armdsp -a command to check the status of your array before making any changes to a VA.

The thing is, it's autoraid. It does almost everything for you. Of course, that also means that it tends to do a lot of things when you least expect it.

When you want to insert a new drive, log into your commandview console, run armdiag -a array-id, and check to see if it's balancing, optimizing or scrubbing, also check for warnings or errors. Just about anything other than those three are bad.

If it's doing one of those three, pop a disk into an empty bay and wait. Drives out of the box are formatted with 512byte sectors. The VA uses 520 byte sectors (512 for data, 32 for checksum), so, it'll take about an hour for a 146GB disk.

Check to see if it's done using the armdsp command.

Never put anything bigger than a 146GB disk into your array. I had a 7400 that had 146GB drives in it, it told us that the biggest disk was a 73GB disk, even though it allocated the space of the 146GB disks. Guess what, if you lose a 146, it only rebuilds 73. This got fixed pretty fast.

If you pull a drive, never put it back in again until the array has finished rebuilding. The array gets really mad if you do this. (I found out once.) It also checks the serial number, so if you move it to a different slot, that doesn't help either.

Oh, and if you are able, always get the firmware updated the the latest rev. Each rev HP adds more problem monitoring entries. Even if it doesn't fix anything, they add more to watch for issues.

If you haven't heard of a bypass_cluster, you are lucky. We had one, it spread like cancer through the array. This was fixed by A120 / HP19.

I've really liked working with these arrays.

Oh, one last thing. HP is end-of-lifing the VA series arrays, April 2006. If you need to buy any new hardware for it, do it before then, otherwise it's off to the used market for you.

Good luck,
Don
Patil P
Occasional Contributor

Re: Procedure to add disk to VA7410

Hello,

1) As per Mr. Don it'll take about an hour for one disk of a 146GB. I am adding 4 *73GB disks Should those be added one by one or I can add at a time?

2) For creation of LUN should I need to wait for array balancing? Or I can create LUN when array is in balancing, optimizing or scrubbing state?

My plan of action is
A) Take backup of OS & Application & database
B) Then run following 4 command as told by HP for one issue
#armmgr -q 15 M/C1.H1 VA_MUM
#armmgr -q 15 M/C1.H2 VA_MUM
#armmgr -q 12 M/C2.H1 VA_MUM
#armmgr -q 12 M/C2.H2 VA_MUM
C) Apply one kernel patch as per HP for some other issue.
As the changes to take effect, VA & Server needs to be restarted, so I will power off the server / VA after normal shutdown & power on VA / server.
D) Then I will add the disk.
E) Creation of LUN,VG, LV etc.

3) I want to know is there any problem on balancing state/old setting of VA, if I Power OFF & Power ON VA?

4) or should I do both this activities separately?

Regards
Patil
Srinivasa_6
Advisor

Re: Procedure to add disk to VA7410

To answer your questions...
1) Adding 4 disks at the same time may not matter that much. Because, if auto-format is enabled, all disks' format are kicked off once their block size is found to be not equal to 520 bytes. Once the auto-format completes, the array will probably kick off balance again to redistribute data optimally among the available drives.
2) LUN creation need not wait for balancing, optimize or scrubbing to complete. These things happen periodically to maintain the array in healthy and optimal state.
3) Just for my understanding, why does the VA need to be rebooted if you apply a patch to your host? I don't think that would be necessary.
4) Actually I've heard somewhere that the state indication updating is awry and a reset of the array would correct that situation. For example: balance might not be running on the array but the status might show the array as balancing. This can be cleared by a reset. I don't know if this is true - had heard about this. AFAIK, there should be no problem rebooting the array while the array is balancing. It would just resume balancing after the reboot.

Srinivas
Don Mallory
Trusted Contributor
Solution

Re: Procedure to add disk to VA7410

Hi Patil,

1) You can add the disks all at once if you add them while the array is powered off.

If you are adding them online, HP recommends that you only add one disk at a time.

2) Agreed with Srinivasa, create / delete LUNs anytime you like. (aside from during a failure, but that will work).

Ahh, the queue thresholds. this value is determined based on a whole bunch of things. Here is the info I received from HP support on this once:

The array has a default of 750. This number is set way to high for most customers systems. The Queue Full Threshold should be set to 2*(# disks in RG) / (controller host ports). However making these changes for MPE or Linux systems may cause the host to have I/O problems. If MPE or Linux systems are connected to the array leave the QFT at the default value.

On a 7400 with 30 disks (16 in RG1 and 14 in RG2), the settings should be:

Controller 1 Queue Full Threshold: 32
Controller 2 Queue Full Threshold: 28

However you cannot set the individual host ports on a VA7400. Set the value to the most appropriate value for both controllers.

On a 7410 with 30 disks (16 in RG1 and 14 in RG2), the settings should be:

Controller 1 Host Port 1 Queue Full Threshold: 16
Controller 1 Host Port 2 Queue Full Threshold: 16
Controller 2 Host Port 1 Queue Full Threshold: 14
Controller 2 Host Port 2 Queue Full Threshold: 14

armmgr -q {Q value} {FRU} {array ID}

Example:
armmgr -q 16 M/C1.H1 test_array_alias
armmgr -q 16 M/C1.H2 test_array_alias
armmgr -q 14 M/C2.H1 test_array_alias
armmgr -q 14 M/C2.H2 test_array_alias



3) We very rarely power cycle our arrays. I'm a little surprised you need to do this unless you are being recommended to by our friends at HP backline support.

If your array shows a Ready state at the VFP (serial console), run "vfpmgr -s shut" and you're down.

This is the safest way to shutdown as it flushes all the buffer caches and commits any remaining data.

If the array is balancing or scrubbing, these are normal states, feel free to do what you want.

4) From #2..
A) - always a good idea.
B) This can really be done anytime with little affect to the host. It's only for performance tweaking.
C) patches.. obviously need a reboot. Still not sure why you are rebooting the array.
D) As said before, if you add all at once while the array is down, life is good. otherwise, just do them one at a time anytime during the day.
E) another operational task.. do it anytime you like.

Don
Patil P
Occasional Contributor

Re: Procedure to add disk to VA7410

Thanks to all, Now I can do my activity with the help of reply received.

Patil p