Unsolved
This post is more than 5 years old
76 Posts
0
1330
Enabling VCM bit on FA ports which are live
We have few FAs where the VCM bit are disabled.
Many AIX hosts are connected to these FAs.
Support updated me that there is no safe way to enable VCM bit.
If a the vcm state is changed while the hosts are accessing disks then there is a chance that data corruption could occur.
For this reason the hosts will need to be temporarily stop i/o on these paths - i.e
We need to stop the db engine, umount the fs, stop the VxVM volumes (or whatever is relevant for your OS).
Has anybody enabled VCM bit on the FAs which are connected to Hosts ?
Any better recommendations where there is no or minimal disruption to the host ?
Many AIX hosts are connected to these FAs.
Support updated me that there is no safe way to enable VCM bit.
If a the vcm state is changed while the hosts are accessing disks then there is a chance that data corruption could occur.
For this reason the hosts will need to be temporarily stop i/o on these paths - i.e
We need to stop the db engine, umount the fs, stop the VxVM volumes (or whatever is relevant for your OS).
Has anybody enabled VCM bit on the FAs which are connected to Hosts ?
Any better recommendations where there is no or minimal disruption to the host ?
dynamox
2 Intern
2 Intern
•
20.4K Posts
1
July 7th, 2008 02:00
MarcT2
131 Posts
1
July 7th, 2008 04:00
AJK3
108 Posts
0
July 7th, 2008 05:00
dynamox
2 Intern
2 Intern
•
20.4K Posts
0
July 7th, 2008 06:00
amrith
76 Posts
0
July 7th, 2008 08:00
These ports are connected to AIX and not to mainframes.
And yes Masking is done on the ports (the host used to see 'extra' luns and hence we found out that the VCM bit is disabled)
Marc,
Powerpath is not installed and it uses the AIX native MPIO.
Can it be done like this:
Mask these all luns to another FA (where vcm is enabled) ,Then unmask from this FA ,Enable VCM and then unmask from the old one.
Will this be a problem ???
xe2sdc
2 Intern
2 Intern
•
2.8K Posts
0
July 15th, 2008 05:00
It went smoothly and worked fine (sometime it happens)
-s-
amrith
76 Posts
0
July 24th, 2008 00:00
I enabled the VCM bit this monday morning and all went fine .
This was the plan I followed :
1. Verify the multi-pathing configuration of each affected host. This is to ensure that the host can remain 'up' while the change takes place. There should be at least 1 path to the storage
2. Bring the port offline on the switch which affects the FA port.
3. Bring Offline the FA port on the particular storage array.
4. Enable the VCM bit
5. Online the FA port.
6. Bring the switch port back online.
7. Ensure that the paths have returned.
xe2sdc
2 Intern
2 Intern
•
2.8K Posts
0
July 24th, 2008 06:00
The same for steps 5 and 6 obviously
I hope you also masked the volumes previously seen by hosts
amrith
76 Posts
0
July 24th, 2008 21:00
null
xe2sdc
2 Intern
2 Intern
•
2.8K Posts
0
August 26th, 2008 03:00
AJK3
108 Posts
0
August 26th, 2008 06:00
MikeMac1
292 Posts
0
August 26th, 2008 06:00
xe2sdc
2 Intern
2 Intern
•
2.8K Posts
0
August 26th, 2008 06:00
MikeMac1
292 Posts
0
August 26th, 2008 07:00
xe2sdc
2 Intern
2 Intern
•
2.8K Posts
0
August 26th, 2008 07:00
HAL is programmed to mimic you, poor human beings
You should know that HAL is a brand new form of AS.
Artificial Intelligence is old. AS is the future