Strange warning after expanding a disk.

[Toine_1]

Hello,

I have expanded a disk on my I64 server.

The EVA4400 crashed (this was a know bug: EVA4400 Controller Terminated During LUN Expansion).

But after a restart of the EVA and I64 server I get the warning below on my disk on the GPT.SYS file. How can I solve this or is this not a problem at all?

$ ana/disk DSA9999:
Analyze/Disk_Structure for _DSA9999: started on 10-JAN-2010 20:26:26.35

%ANALDISK-W-MULTALLOC, file (11,11,0) [000000]GPT.SYS;1
multiply allocated blocks
VBN 97 to 144
LBN 419430352 to 419430399, RVN 1
%ANALDISK-W-MULTALLOC, file (11,11,0) [000000]GPT.SYS;1
multiply allocated blocks
VBN 145 to 192
LBN 419430352 to 419430399, RVN 1


/Toine

[John Gillings]

Toine,

Interesting! This means the range of physical blocks LBN 419430352 to 419430399 is allocated twice. In this case to diferent (but contiguous) places in the same file.

The usual fix for MULTALLOC is to make copies of each file involved and delete the originals.

However, GPT.SYS is a system file. It has something to do with faking out a PC like file system on an OpenVMS disk. It depends on physical locations, so you can't just delete it.

If it's not a bootable disk, it probably doesn't matter. A BACKUP/IMAGE might help eliminate the error. Other than that, I'm sure Hoff will have something to say about it.

[Toine_1]

John,

Thanks for the answer.

It isn't a system disk. Can I just delete that file ?
Is this file GPT.SYS also used on non system disks?

/Toine

[John Gillings]

Toine,

Definitely don't delete GPT.SYS! (I doubt the file system would let you). For now I wouldn't lose any sleep about it.

Wait to see what Hoff says.

[Hoff · Accepted solution]

Short answer: BACKUP /IMAGE your data off this (corrupt) disk volume, reinitialize it, and reload it.

In parallel, verify your patch level on OpenVMS and you should be on V8.3-1H1 with some reasonably recent UPDATE kit, and ensure your EVA firmware and your server are both sufficiently fresh.

GPT.SYS file is a protective file; the blocks upon which this file is placed are read and sometimes written at the console level, and the OpenVMS storage allocation and file system view here is merely a reservation for this disk storage.

HBVS use here has a second consideration, as should the console detect a corruption and decide to correct the corruption and write to the structure because it has detected a GPT corruption, the write is NOT coordinated with the other HBVS volumes. Some EFI versions reportedly prompt for permission to perform the write, but I've not seen that.

If you're interested in (more) details of the low-level disk structures including GPT.SYS (I designed this stuff), those details are posted on my web site.

Regardless of any technical interest in the disk volume structures, this disk structure is corrupted. Something corrupted system files. What else got corrupted is an open question.

Get your data off this disk. Initialize it. And reload it.

[Toine_1]

Thank you Hoff,

I don't think any other files are corrupt.
ANA/disk/read didn't give any other warnings.
I will do a backup/image and init the disk and do a restore so I'm sure it is OK.

The EVA is on latest firmware and the I64 server has almost the latest patches installed.

NVP$ prod sho hist *update*
------------------------------------ ----------- ----------- --- -----------
PRODUCT KIT TYPE OPERATION VAL DATE
------------------------------------ ----------- ----------- --- -----------
HP I64VMS VMS831H1I_UPDATE V6.0 Patch Install Val 28-SEP-2009


Thanks you for the info.

/Toine