Thursday, May 26, 2005
Our Standard Symmetrix DMX Setup
- 000 Volume Logics database (8mb)
Contains information as to what host has access to which meta devices. Should be read only to everyone except systems with Solutions enabler installed.
- 001-00F Gate Keepers (3mb)
All EMC software (TimeFinder, SRDF, SDR, etc.) run IN the Symmetrix, so servers (with SYMCLI or ECC) only issue commands to the Symmetrix and Symmetrix perform the job internally. The problem is that Symmetrix boxes can not assign memory for software, only devices, so (cleverly) EMC creates tiny devices of 6 cilinders (aprox. 2,88 MB) in order to give the requiered memory IN the Symmetrix, so symm box can receive the commands from symcli/ECC through gatekeeper devices (to the symmetrix memory) and perform it.
- 010-019 Quarum Devices (range from 70-668mb)
Quarum devices for windows clustering.
- 01A-01B Symmetrix File systems (2878mb)
For use by EMC CE only
- 01C-01F Sym Optimizer Devices (8632mb)
Used by Symm Optimizer to move hot spots in the Sym
- 020-### Meta devices pathed out to hosts in increments of 8.43G
Actual meta devices as seen by hosts.
Command example:
symdev -sid #### list more
Symmetrix ID: ###########
Dir | Device | ||||||
Device Name | SA:P | DA:IT | Config | Attribute | State | Size (MB) | |
0000 | /dev/rdsk/c4t32d0s2 | 03A:0 | 01A:C2 | 2-Way Mir | N/Grp'd | VCM WD | 8 |
0001 | /dev/rdsk/c4t32d1s2 | 03A:0 | 01A:C1 | 2-Way Mir | N/Grp'd | RW | 3 |
0002 | /dev/rdsk/c4t32d2s2 | 03A:0 | 02B:C1 | 2-Way Mir | N/Grp'd | RW | 3 |
0003 | Not Visible | ***:* | 07A:C1 | 2-Way Mir | N/Grp'd | RW | 3 |
0004 | Not Visible | ***:* | 08B:C1 | 2-Way Mir | N/Grp'd | RW | 3 |
0005 | Not Visible | ***:* | 01B:C1 | 2-Way Mir | N/Grp'd | RW | 3 |
0006 | Not Visible | ***:* | 02A:C1 | 2-Way Mir | N/Grp'd | RW | 3 |
0007 | Not Visible | ***:* | 07B:C0 | 2-Way Mir | N/Grp'd | RW | 3 |
0008 | Not Visible | ***:* | 08A:C0 | 2-Way Mir | N/Grp'd | RW | 3 |
0009 | Not Visible | ***:* | 16B:D0 | 2-Way Mir | N/Grp'd | RW | 3 |
000A | Not Visible | ***:* | 15A:D0 | 2-Way Mir | N/Grp'd | RW | 3 |
000B | Not Visible | ***:* | 10B:D0 | 2-Way Mir | N/Grp'd | RW | 3 |
000C | Not Visible | ***:* | 09A:D0 | 2-Way Mir | N/Grp'd | RW | 3 |
000D | Not Visible | ???:? | 16A:D0 | 2-Way Mir | N/Grp'd | RW | 3 |
000E | Not Visible | ???:? | 15B:D0 | 2-Way Mir | N/Grp'd | RW | 3 |
000F | Not Visible | ???:? | 10A:D0 | 2-Way Mir | N/Grp'd | RW | 3 |
0010 | Not Visible | ***:* | 08A:DD | 2-Way Mir | N/Grp'd | RW | 668 |
0011 | Not Visible | ***:* | 09B:DD | 2-Way Mir | N/Grp'd | RW | 668 |
0012 | Not Visible | ***:* | 08A:DD | 2-Way Mir | N/Grp'd | RW | 668 |
0013 | Not Visible | ***:* | 09B:DD | 2-Way Mir | N/Grp'd | RW | 668 |
0014 | Not Visible | ***:* | 01A:C0 | 2-Way Mir | N/Grp'd | RW | 70 |
0015 | Not Visible | ???:? | 02B:C0 | 2-Way Mir | N/Grp'd | RW | 70 |
0016 | Not Visible | ???:? | 07A:C0 | 2-Way Mir | N/Grp'd | RW | 70 |
0017 | Not Visible | ???:? | 08B:C0 | 2-Way Mir | N/Grp'd | RW | 70 |
0018 | Not Visible | ???:? | 01B:C0 | 2-Way Mir | N/Grp'd | RW | 70 |
0019 | Not Visible | ???:? | 02A:C0 | 2-Way Mir | N/Grp'd | RW | 70 |
001A | Not Visible | ???:? | 08A:DD | 2-Way Mir | N/A | (FS) RW | 2878 |
001B | Not Visible | ???:? | 09B:DD | 2-Way Mir | N/A | (FS) RW | 2878 |
001C | Not Visible | ???:? | 01B:DD | DRV | N/Grp'd | RW | 8632 |
001D | Not Visible | ???:? | 16A:DD | DRV | N/Grp'd | RW | 8632 |
001E | Not Visible | ???:? | 07B:DD | DRV | N/Grp'd | RW | 8632 |
001F | Not Visible | ???:? | 10A:DD | DRV | N/Grp'd | RW | 8632 |
0020 | Not Visible | ***:* | 01A:C0 | 2-Way Mir | N/Grp'd | (M) RW | 60421 |
Monday, May 16, 2005
Fixing a non-refresh problem on ECC
2. Inside ECC stop the Symmetrix agent on all of the store boxes.
3. Inside ECC start the Symmetrix agent on all of the store boxes, starting the sym agent on the box which you want to act as primary first.
This will cause a database refresh and hopefully clear the problem.
Finding HBA WWNs
1. Run LPUTIL to gather the WWN of the HBA (note this is only for Emulex cards):
/usr/sbin/lpfc/lputil
2. Select 2. Adapter Information > 4. IEEE Address > Card in question.
IEEE Address for Adapter 2:
[10000000] [999999999]
HPUX:
1. Run an IOSCAN to locate the fiber card device files:
ioscan -funC fc
2. Using a device from above run the FCMSUTIL command to get the card information:
/opt/fcms/bin/fcmsutil /dev/td0
Friday, April 15, 2005
Creating Meta's in ECC
The first step in giving a host access to EMC disks is to actually create the metas themselves. The metas are logical collection of standars that are set up to look like a single device to the actual hosts. Always use/create RAID5 devices as concatenated, RAID1 devices as striped.
1. In ECC select:
- Storgare systems
- Symmetrix Arrays
- Right click on the appropiate Symm
- Select configure
- Select Meta Device Configuration
2. On the left side select the appropiate Unmapped Devices
- BCV Devices are for Business continuous volumes
- Standards are for regular volumes
- For RAID5 volumes select four contiguous standards/BCV (this is what we do, this isn't what everyone does)
3. On the right side select the correct type of Meta Device you wish to create. Always use/create RAID5 devices as concatenated metas , RAID1 devices as striped metas.
4. Push the Create Meta button!
Note: The newly created Meta Devices will be created using the first standard/BCV in the meta as it's identifier. This number, the meta head, is not the LUN Device name/number.
Adding Devices to a FA ECC
1. In ECC select:
- Storgare systems
- Symmetrix Arrays
- Right click on the appropiate Symm
- Select configure
- Select SRD Device Mapping
2. On the left side select the needed Meta Devices
BCV Devices are for Business continuous volumes
Standards are for regular volumes
3. On the right side select the correct FA and port address to map the devices down. FAports should be select using the rule of seventeen (i.e. 7+10, 8+9,...)
4. Push the Continue button and then edit the command.
5. Changed the automatically assigned LUN numbers to the desired LUN number (we start our LUNs at address 20 or if there are already LUNs assigned to the FA then we use the next available LUN number).
6. Press the continue button and you are done!
HBA access via ECC
1. In ECC select:
Hosts
Right click on the appropiate Hostname
Select Masking
Select Modify Masking Configurations
2. In the new window, select the desired HBA in the top left window pane.
3. On the top right window pane select the correct FA and port address to mask the devices down. FAports should be select using the rule of seventeen (i.e. 7+10, 8+9,...)
4. Push the Show Devices button.
5. Select the desired devices in the window bellow and the Grant Access. Push continue to commit the changes.
HBA access via Store Command Line
1. Login to the ECC store NT box.
2. Open the command prompt and type :
symmaskdb list database more
3. Verify that the devices you are pathing are free on the list.
4. Grant access to the devices in question with the following command:
symmask -sid "Symm-Serial Number" -wwn "WWN Number of HBA" -dir "FA" -p "FA Port Number" add devs "Device range"
example:
symmask -sid #### -wwn 10000000c9fake -dir 9c -p 0 add devs 62a:78F
5. Verify that the access has been granted:
symmaskdb list database more