Sun Cluster/Storage question

Hi Bros,
I am a newbie to this cluster environment...my workplace is a mess of Sun solaris server, oracle database, Cisco switch, router, MS Server, Windows PC..so and so..now my boss ask me to come out with a procedure for maintenance of storage array in our sun cluster environment..I read few articles from Oracle but really hard to figure it out..anyone have a procedure for Storage array maintenance like shutdown of storage array, HDD changing etc...Thanks in advance

we have two nodes (sun V490) with Oracle 9i database.. I know how to safely shutdown, restart, change HDD for this two nodes but I didn't really touch this storage array before in sun cluster.. Hope someone can help...

well, it would be good to know what "this" storage is... and how it is connected... and and and...

Hi SoulMan,

Thanks for the quick response.. I am not sure how to explain..this two storage arrays using sun storedge consist of 6 harddidsks each, connected through SCSI cable and we have two resource groups in quorum...

look, this is not helpfull. see it from your side, if you were given the infos you give us, would you be able to come up with a procedure to replace a disk?

if you don't know the modell of the storage, a photo might be helpfull. also the configuration of the cluster and storage is helpfull. here are some infos we might need to help you:

• solaris version
• cluster version
• config of cluster (scstat/scdidadm/...)
• ... (means all information that is helpfull)

also, you should have a support contract for the system? if it is an important system there should be!

you can find all the documentation from oracle at Oracle Documentation for more infos about the products to get you started!

Yes I know sorry for the trouble..I am trying to upload scstat image

---------- Post updated at 03:10 AM ---------- Previous update was at 02:46 AM ----------

SunFire V490 Servers(2 nodes)
Sun Storedge 3300
Cluster version 3.1

just put XXX's for sensitive data like IP addresses and so on... we can only help if you post here for the benefit of all members.
if you can't post here, you should use your (hopefully you have it) support contract and open a servicerequest at http://support.oracle.com

Hi All,

Thanks for the advice, We have a contract but as per the contract we only have 6 expert visit per year and also they charge $2000 for every visit. So my boss looking for alternative and ask me to come out with the procedure to first line maintenance such as HDD changing, proper shutdown/restart in case of cluster error...

OK, there is no such support contract with 6 visits a year. Oracle only has a "premium support" and that is not limited.

Next is, your output does not show any problems. If a disk is failed (regarding to your first post) and the disk is inside a SE3300 box (which is JBOD) all disks will be visible to your systems. there has to be some sort of volumemanagement involved to create cluster manageable volumes (slvm, zfs or vxvm). you need to use the product specific procedure to replace a failed disk in such a volume.

Hi Mr. Duke,

That contract is not with Oracle but it with a external IT support firm..That is why..

Now all working normal but I have some doubts to clear

1. What are the symptoms I mean common faults we look for in this cluster and storage array from the out put
2. How to gracefully shutdown and restart this storage array after disconnected from nodes (Both nodes need to down first am I right? I am able to shutdown and restart both nodes without any issue)

Thanks again for the advice

1. you have to look for errors, like offline resources which should be online. check iostat -eE for errors on the disks...

2. there are only two on/off switches for each power supply on the storage. to shutdown the storage, you need to switch BOTH ps to off (after shutting down the cluster). you can start both nodes in non-cluster mode by entering boot -x from ok prompt.

Hi Mr. Duke,
Just to confirm you are saying it's okay to power down the storage by pressing PS button after I shutdown two nodes properly by # scshutdown �y �g0 ?

Need to power up storage before boot up nodes even though entering boot -x right ?

We done it before power down and power up the storage but my boss is saying there should have some steps to shutdown the storage after shutdown 2 nodes by a laptop connecting directly to storage array

Is there any such trouble-shooting steps that is what I am looking for...but I read all oracle documents unable to find... my bad...please help

Thanks a lot

Contents

I find it useful...hope able to work around it..Thanks