| Blogs | Classifieds | Downloads | FlashChat | Gallery | Googlemap | Invite Friends | Links | Projects | Reviews | Wiki |
| |||||||||
|
#1
| ||||
| ||||
I need to configure two lpars to read/write to the same SAN disk and would like some tips. Some background: we have 2 p570s for application servers and a DS4800 storage system. The production database is on frame B and the development db is on frame A. I am building a new partition on frame A which will be a standby server for the production server. The plan is that the live and standby servers will be able to read and write to the same SAN disk (though at different times) for availability in case the production server goes down with a non-disk related problem. (...a disk failure will affect both servers. A separate DR solution is planned but the Technical Architect wants this standby server as well as the DR plan. DR will only be invoked in case of disk failure) I will be adding the standby server to the same storage partition on the SAN (in order to assign the same disks to both partitions) but beyond this I am not sure how to get the standby server to read and write to the file systems created by the live server. If you have done this before or have ideas about this type of configuration I would appreciate any tips that you have. For example, 1) would I need to backup the volume group config information on the live box and restore them on the standby server ? 2) How do I get Oracle software to run on the new partition without re-installing since the directories and volume group structures are exactly the same? Many thanks Jisco
__________________ "Love never fails...do unto others as you would have them do to you". To view links or images in signatures your post count must be 10 or greater. You currently have 0 posts. |
|
#3
| ||||
| ||||
Hi Billgncal No. There is no failover software solution being considered. The impression here is that SAN facilitates disk sharing so two lpars should be able to read/write to the same disk without any problems. HACMP will be considered too expensive a solution by miles. The requirements for this have complicated a little further as they now want this lpar to have to have another set of disks assigned to it for a test databse AS WELL AS sharing the production server disks. i.e that they would like to be able to completely trash the test db and simply mount the production filesystems and and continue seamlessly or even mount both simultaneously. Cheers
__________________ "Love never fails...do unto others as you would have them do to you". To view links or images in signatures your post count must be 10 or greater. You currently have 0 posts. |
|
#4
| ||||
| ||||
It is true that you can set up the SAN so that two server can see the disks. But only one server can be using them at a time. You only need to varyon the vg on the server but not at the same time. Is this what you are refering to? |
|
#5
| ||||
| ||||
Hi there HACMP isn't actually that expensive. Once configured, will provide you with a fully automated fallover to your standby server. When initially configuring the HACMP shared volume group you would first perform the following. Use this to resolve your issue 1. Create a VG and specify a unique major number 2. create your lv's and filesystems 3. varyoff the volume group 4. On your standby node, run cfgmgr to bring your disks in. lspv, should show matching pv id's on the primary and standy. If not, run chdev -l pv=yes hdisk# for each disk. 5. Import the volume group using the same major number as step1 6. mount the file systems You obviously need to automate this via a script should your primary machine fail. Without HACMP, this could get a bit messy HTH Steve
__________________ Take your time and think things through !! To view links or images in signatures your post count must be 10 or greater. You currently have 0 posts. |
|
#6
| ||||
| ||||
Hi I would install the oracle binaries on each server. You should only have the datafiles containing your database on the shared disk Steve
__________________ Take your time and think things through !! To view links or images in signatures your post count must be 10 or greater. You currently have 0 posts. |
|
#7
| ||||
| ||||
posted in error- withdrawn
__________________ "Love never fails...do unto others as you would have them do to you". To view links or images in signatures your post count must be 10 or greater. You currently have 0 posts. |
|
#8
| ||||
| ||||
Hi billgncal and steevojb, Thanks a lot. Steevojb, many thanks for the detailed procedure. Yes, this is exactly what I am looking to do. I would like to clarify one thing from your procedure. The required VGs are already live on the production server so I guess step 5 would mean importing them into the standby server using the major numbers as they currently exist on the live server? Best regards Jisco
__________________ "Love never fails...do unto others as you would have them do to you". To view links or images in signatures your post count must be 10 or greater. You currently have 0 posts. |
|
#9
| ||||
| ||||
Hi There Yes, you need to importvg using the same major numbers. You may be able to do this online, although I would not recommend it. Have a look at importvg -L If you have spare LUNS, I would test everything first to ensure you do not get any unexpected results HTH Steve
__________________ Take your time and think things through !! To view links or images in signatures your post count must be 10 or greater. You currently have 0 posts. |
|
#10
| ||||
| ||||
I have a question relating to this thread, I have a p5 520 and will run AIX5.3. There are 2 fiber channel cards. It will be hooked to an HP SAN. Will I need to have additional software to handle the dual pathing? |
![]() |
| Bookmarks |
| Currently Active Users Viewing This Thread: 1 (0 members and 1 guests) | |
| Thread Tools | |
| |
Similar Threads | ||||
| Thread | Thread Starter | Forum | Replies | Last Post |
| VIOS and CD drive sharing | mihha | IBM PowerVM Editions | 2 | May 24th, 2007 02:02 |
| SSA reset disk | silvia | AIX for POWER Systems | 17 | February 19th, 2007 15:15 |
| VIO Disk Performance | gbarth | IBM PowerVM Editions | 7 | January 8th, 2007 15:42 |
| Quit disk reserve from rootvg | flama | AIX for POWER Systems | 2 | September 30th, 2006 08:17 |
| Disk I/O Benchmarking | pedz | AIX for POWER Systems | 10 | August 18th, 2006 12:35 |