- Community Home
- >
- Servers and Operating Systems
- >
- Operating Systems
- >
- Operating System - HP-UX
- >
- Re: HELP!!! SYSTEM DOWN-- MAY LOSE DATA
Categories
Company
Local Language
Forums
Discussions
Forums
- Data Protection and Retention
- Entry Storage Systems
- Legacy
- Midrange and Enterprise Storage
- Storage Networking
- HPE Nimble Storage
Discussions
Discussions
Discussions
Forums
Forums
Discussions
Discussion Boards
Discussion Boards
Discussion Boards
Discussion Boards
- BladeSystem Infrastructure and Application Solutions
- Appliance Servers
- Alpha Servers
- BackOffice Products
- Internet Products
- HPE 9000 and HPE e3000 Servers
- Networking
- Netservers
- Secure OS Software for Linux
- Server Management (Insight Manager 7)
- Windows Server 2003
- Operating System - Tru64 Unix
- ProLiant Deployment and Provisioning
- Linux-Based Community / Regional
- Microsoft System Center Integration
Discussion Boards
Discussion Boards
Discussion Boards
Discussion Boards
Discussion Boards
Discussion Boards
Discussion Boards
Discussion Boards
Discussion Boards
Discussion Boards
Discussion Boards
Discussion Boards
Discussion Boards
Discussion Boards
Discussion Boards
Discussion Boards
Discussion Boards
Discussion Boards
Discussion Boards
Discussion Boards
Discussion Boards
Discussion Boards
Community
Resources
Forums
Blogs
- Subscribe to RSS Feed
- Mark Topic as New
- Mark Topic as Read
- Float this Topic for Current User
- Bookmark
- Subscribe
- Printer Friendly Page
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
тАО03-07-2001 04:54 PM
тАО03-07-2001 04:54 PM
Background...system lost power tonight, and upon rebooting, vg02 did not activate.
vg02 consists of 2 hard disks, with mirroring in effect, /dev/dsk/c3t2d0 and /dev/dsk/c7t2d0.
On bootup, /dev/dsk/c7t2d0 does not respond to any command, including dd, diskinfo, pvdisplay, vgcfgrestore, etc., but does show up as 'CLAIMED' on ioscan.
vgchange would not work on vg02 due to this messed up disk. HP had me do a no-quorum vgchange as follows: vgchange -a y -q n /dev/vg02, which took a while, then after giving an error about the bad disk, said vg02 was activated successfully. Now, my lvdisplay of /dev/vg02/uvol2 shows the following:
lvdisplay -v /dev/vg02/uvol2--- Distribution of logical volume ---
PV Name LE on PV PE on PV
/dev/dsk/c3t2d0 508 508
--- Logical extents ---
LE PV1 PE1 Status 1 PV2 PE2 Status 2
00 /.../c3t2d0 00 stale ??? 00 current
01 /.../c3t2d0 01 stale ??? 01 current
Note: the good disk (c3t2d0) shows stale, while the bad disk, which it can't even name (c7t2d0) shows current--for all extents.
lvdisplay -v -k /dev/vg02/uvol2 shows:
--- Logical extents ---
LE PV1 PE1 Status 1 PV2 PE2 Status 2
00 0 00 stale 1 00 current
01 0 01 stale 1 01 current
Is there ANY way to salvage these 'stale' extents on the good disk? I can't afford to lose data, and what's the point of having Mirrordisk and mirroring configured if one bad hard disk can wipe out my data anyway? Thanks for any help!!!!
Jeff
Solved! Go to Solution.
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
тАО03-07-2001 05:10 PM
тАО03-07-2001 05:10 PM
SolutionOutput is a little strange. Is the file system mounted and usable. I ask because the output seems to show that LVM thinks one disk has stale extents and the other disk it cannot even recognize. This is not a normal condition with mirroring. For example, if this were hot-pluggable and you were absolutely sure that c3t2d0 is the bad disk then you would reduce your mirrors:
# lvreduce -A n -m 0 /dev/vg02/uvol2 /dev/dsk/c3t2d0 (using the disk that is bad).
You can run lvlnboot -R /dev/vg02 and see if that will correct the readout on the ????
If c3t2d0 is actually the bad disk and you do this then you've lost your data.
If the disk with the ??? is the bad one then you're hosed - no way to recover the data.
Once you've reduced the LVM data off the bad disk then if hot-pluggable the disk can be replaced. Then you need to rebuild the disk from scratch and c3t2d0 will become the mirror.
Another option is to halt the system replace c3t2d0. Go to single-user mode. Run:
vcfgrestore -n vg02 /dev/rdsk/c3t2d0
vgchange -a y vg02
If good, then reboot to multi-user mode.
Good Luck.
Peggy
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
тАО03-07-2001 05:20 PM
тАО03-07-2001 05:20 PM
Re: HELP!!! SYSTEM DOWN-- MAY LOSE DATA
I read your note too quickly and now see that you already said that c7t2d0 is the disk that does not respond and the other disk has the stale extents. There was another posting where someone wanted to get there data back. It's a shot in the dark but it might work. I'll see if I can find that post. Sorry for the bad luck.
Peg
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
тАО03-07-2001 05:41 PM
тАО03-07-2001 05:41 PM
Re: HELP!!! SYSTEM DOWN-- MAY LOSE DATA
I found the posting that might help. It is:
http://forums.itrc.hp.com/cm/QuestionAnswer/1,1150,0x7cc06af52b04d5118fef0090279cd0f9,00.html
Just in case you cannot get to it I will copy it here:
Problem:
We replaced a disk with a larger one some time ago. We did not transfer some old information to the new disk, but now the users want it back and our backup tapes have aged out. The disk has had the LV removed using SAM but nothing else (i.e. VG still defined, etc). Is there any way to nondestructively re-define the LV definition to allow access to the data on the disk?
Thanks
Answer by Andy Monks (HP)
yes there is, however it might take a few attempts.
If it was the last lvol in the volume group, just doing a lvcreate of a size greater or equal to was it was should do.
If it was in the middle, then you'll have to know the size, but as long as no other lvremoves have been done, again a lvcreate (of the exact size) should work.
none of the lvm command will actually touch the data, so it doesn't matter how many times you attempt this.
What you quite often have to do, is create a fake lvol just so that you can get lvm to create a new lvol where the old one was (this is only if you've deleted a 2 or move lvols).
This worked for the person requesting the info.
You would have to adapt this to your situation.
I think I would get a complete listing of the lvols and sizes, etc. on the disk.
Reduce the lvol from c7t2d0. Example:
lvreduce -k -m 0 /dev/vg02/uvol2 1
All that you have now are the stale extents on the disk c3t2d0. lvremove that lvol and then re-create it. Since it won't have a mirror it should come up as current. If you get it exactly in the position it was before, you might recovery your data.
Again, Good Luck.
Let me know what happens.
Peggy
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
тАО03-07-2001 06:40 PM
тАО03-07-2001 06:40 PM
Re: HELP!!! SYSTEM DOWN-- MAY LOSE DATA
Thank you so much for all the info you've given me...HP's techs are working on the case right now...I like your idea about lvremoving & lvcreating the lvol to try to salvage the data--there's only 1 lvol per disk. Before HP goes further, would you recommend that I dd the entire good (stale) disk out for safe keeping? Thanks again.
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
тАО03-07-2001 07:01 PM
тАО03-07-2001 07:01 PM
Re: HELP!!! SYSTEM DOWN-- MAY LOSE DATA
-Jeff
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
тАО03-07-2001 07:25 PM
тАО03-07-2001 07:25 PM
Re: HELP!!! SYSTEM DOWN-- MAY LOSE DATA
What a great idea to dd the disk - I hadn't thought of that one. At least you didn't need it. So, did you do a vgcfgrestore on the replaced disk? Pretty amazing. Normally mirror software works great but every once in awhile you see some bizarre stuff. I am really glad it worked out without you having to do any extreme measures....
and thanks for all the points (you didn't have to do that!)
Peggy
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
тАО03-08-2001 10:27 AM
тАО03-08-2001 10:27 AM
Re: HELP!!! SYSTEM DOWN-- MAY LOSE DATA
This one is unusual, because the survivor was marked 'stale', I think this means some information didn't get synched to the disk, possibly you are using async IO?
Doing an ioscan -f should have normalized the information for the dead disk, it should be able to recognize NO_HW there, not sure if this was tried.
Once this is done, the only way I know to clear up the ??? problem is to do vgreduce -f followed by vgscan. If LVM and the kernel are still out of sync, I think rmsf -a on the dead disk devname is also helpful, then another vgreduce -f / vgscan might have worked.
This said with the quailfier that the 'stale' doesn't match our experience, but apparently you were able to recover by other means and any corruption was repairable. I am assuming support has looked over your OS and patches. Good luck.
- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Permalink
- Report Inappropriate Content
тАО03-09-2001 11:56 AM
тАО03-09-2001 11:56 AM
Re: HELP!!! SYSTEM DOWN-- MAY LOSE DATA
Dave, thanks also for your help...I did try several lvm commands, i.e. lvreduce & vgreduce, but it wouldn't let me do either because of the state the disk was in. Also, I did run ioscan -f, several times, and each time it showed the disk in question as if there were nothing wrong, i.e. CLAIMED, even though I couldn't do anything else with it.
Thanks again to both of you...
Jeff