disk group configuration different after delete dev
On my DMX3 I had nine disk groups from initial setup, numbered 0 through 8. Each disk group contained either 300GB 10k drives or 146GB 15k drives, and a couple disk groups contained some of each. As I needed to create devices I would specify the disk group I wanted to use within the symconfigure command.
Recently I have cleaned up somewhat, deleting about 100 unmapped devices not being used. I have been using the symconfigure delete dev command to do this.
I notice now that I no longer have devices spread across all nine disk groups, but rather all devices show up in disk group 0 only, and disk groups 1 through 8 are empty. Both symmdisk list -disk_group # and symmdev list -disk_group # give similar results: only disk group 0 returns entries and the other disk groups appear empty.
How is that possible? I thought only a bin change could alter disk groups, and I have not had a bin change in over a year. Can disks be "condensed" into a single group? Is the Symmetrix doing some optimizing? Should I even care?
Previously I might have seen some 520 drives in a symdisk list -by_diskgroup output, spread out over a total of nine disk groups. Now I see that number appearing only under disk group 0 and all other disk groups show no drives.
Some of the devices deleted as part of cleanup activities included a large number of gatekeeper devices, 111 in all. The gatekeepers were not mapped and, as I assumed, not needed or being used. I'm wondering if the removal of the gatekeepers could have caused this.
Previously I might have seen some 520 drives in a symdisk list -by_diskgroup output, spread out over a total of nine disk groups. Now I see that number appearing only under disk group 0 and all other disk groups show no drives.
that's weird, you don't see symmetrix devices or you don't the actual physical spindles ? I would expect them to show up something like this, 0 hypers per drive. Does running "symcfg discover" change anything ?
DF-5A 05A D F SEAGATE C300LPF 0 286102 286102 286102 DF-5B 05B D E SEAGATE C300LPF 0 286102 286102 286102 DF-12B 12B D F SEAGATE C300LPF 0 286102 286102 286102 DF-6B 06B C E SEAGATE C300LPF 0 286102 286102 286102 DF-5C 05C D F SEAGATE C300LPF 0 286102 286102 286102 DF-11C 11C C E SEAGATE C300LPF 0 286102 286102 286102 DF-12D 12D C E SEAGATE C300LPF 0 286102 286102 286102 DF-12D 12D D F SEAGATE C300LPF 0 286102 286102 286102
Problem appears to be resolved and it looks like it was an issue with my host running SE.
Solutions Enabler had been upgraded recently. I had done most of my cleanup work prior to the upgrade and I was basing my knowledge of how the disks were spread across nine groups from that time frame. The layout showing only diskgroup 0 was post-upgrade of SE. I started suspecting a host issue when an EMC analyst dialing into the box said he could see all nine disk groups. I was concerned because I had tried creating a single GK device as a test, using disk group 0 (as that was the only one I could see) and the command failed for lack of space. Although I could see all 520-odd devices under disk group 0, it's probably true that the disks belonging to disk group 0 really were all full. Suspecting the host, I then tried to create the GK in disk group 4 which should have had ample space, and the command succeeded. Following that, I could again see all nine disk groups with symdisk list -by_diskgroup.
I suspect that creating the GK must have refreshed some local file or db on the ECC host I'm using. Oddly, I had tried symcfg configure twice but that didn't help. I have heard there is some configuration file on the host that sometimes needs to be manually deleted before running symcfg configure, but I'm not sure where that file is.
brad12341
207 Posts
349
0
Posted February 12th, 2010 15:00