topic Re: VA7410 space reservation was desired, but the space left in Disk Enclosures

VA7410 space reservation was desired, but the space left

Lemeshko — Tue, 20 Jun 2006 05:32:51 GMT

Hi ALL,

We have RP5405*2(Serviceguard), VA7410, 15*74Gb.

One of 15 disks has yellow indication.

In syslog.log we have a lot of such errors:

vmunix: SCSI: Request Timeout; Abort -- lbolt: 1348018328, dev: bc032000, io_id: 303a05f

armdsp -a - see attachment.

All disks have status GOOD, but several from them has warning message :

" Space is reserved on this disk for subsystem metadata and
may be a map disk."

At the bottom of armdsp report we can see such message:

"WARNING: The host indicated that hot spare space reservation was desired, but the space left unallocated was insufficient to provide the reservation."

I read information about this error and
advice was - add more disks. :)

Any help would be very appreciated,
Igor.

Re: VA7410 space reservation was desired, but the space left

Owen_15 — Thu, 19 Jul 2007 11:15:14 GMT

Hi Igor,

You have an error with disk M/D15.

Disk at M/D15 is not included at the moment because it was previously used in another array. Did you replace this recently? If you did, you need to tell the array to include this disk and destroy whatever data is on it. This is probably the disk that is showing the amber light.

This will give 73gb back to RG1 and then the status of the array should change from warning to normal again, and the message should dissapear.

I would recommend that you leave some space in the unallocated pool next time you reconfigure the array. This leaves space for the array to perform it's house keeping tasks. I can see that you've got the array in RAID1+0 and that you just have single big luns in each RG, so it may not be possible for you to reduce the size of your luns. In that case, you just need to ensure you monitor the array and respond to any faults promptly.

Regards
Owen

Re: VA7410 space reservation was desired, but the space left

tkc — Fri, 20 Jul 2007 00:04:22 GMT

The error in syslog "vmunix: SCSI: Request Timeout; Abort -- lbolt: 1348018328, dev: bc032000, io_id: 303a05f" was due to the VA7410 array slow in responding to I/O request and probably due to lesser space in the array. Putting back the disk to give back some space to the RG should improve the situation.