VMware EMC SAN storage groups minimum size available space
Hello Everyone,
A general question about VMware 6.0 best practice re. total storage group size available.
We have an EMC VNXe3150 SAN with 4 Storage groups all around 1.8 TB of usable space. Some of them have a lot of space left whilst others are down to what I will call bare minimum, e.g. 70 GB and 100 GB out of the total 1.8 TB. I did not use the rest of the space as I figured it is best to leave a bit for the system, but that is based on my experience with physical servers. I did get a warning from the VMware about space running out and so I left it. Should I leave more? Oh and just to clarify the 3 ESXi hosts have disks as well (we do not host the Linux o/s via USB or SD card, rather built in disk) and they are around 130 GB and 80% free so I am talking only about the actual storage available from the attached SAN to the VM's through the 3 hosts. Finally we always Thick provision (Eager Zero) so all space needed is always assigned, i.e. I don't over provision so the space left over should never be needed by a VM for space.Thanks in advance!
A general question about VMware 6.0 best practice re. total storage group size available.
We have an EMC VNXe3150 SAN with 4 Storage groups all around 1.8 TB of usable space. Some of them have a lot of space left whilst others are down to what I will call bare minimum, e.g. 70 GB and 100 GB out of the total 1.8 TB. I did not use the rest of the space as I figured it is best to leave a bit for the system, but that is based on my experience with physical servers. I did get a warning from the VMware about space running out and so I left it. Should I leave more? Oh and just to clarify the 3 ESXi hosts have disks as well (we do not host the Linux o/s via USB or SD card, rather built in disk) and they are around 130 GB and 80% free so I am talking only about the actual storage available from the attached SAN to the VM's through the 3 hosts. Finally we always Thick provision (Eager Zero) so all space needed is always assigned, i.e. I don't over provision so the space left over should never be needed by a VM for space.Thanks in advance!
ASKER
Sorry, maybe my explanation was not clear. I have 3 hosts that each run VMware from locally installed disks (130GB each and 80% free). I also have a SAN with a total capacity of 7.96 TB and 3.10 TB available split into 4 storage groups in one cluster (see below)
Datastore0 -> Capacity 1.80 TB, Free 1.00 TB
Datastore1 -> Capacity 1.80 TB, Free 71 GB
Datastore2 -> Capacity 1.99 TB, Free 105 GB
Datastore3 -> Capacity 1.99 TB, Free 1.55 TB
Note that I have space on two of the storage groups 0 and 3, but am down to 71 GB and 105 GB FREE on storage groups 1 and 2. I typically "migrate only the virtual machines to another host" for maintenance etc. but not their storage groups since the SAN does not go down for ESX updates and when the entire SAN needs to be rebooted then migrating storage would not help as we have only 1 SAN, i.e. I don't normally have to "migrate the virtual machines' storage to a compatible datastore or datastore cluster" nor have I had to "migrate the virtual machines to a specific host or cluster and their storage to a specific datastore or datastore cluster" but I just tried it and I was able to migrate a vm's storage only from Datastore1 (free 71 GB before migration) to Datastore0 (free 1 TB before migration) without issues and the resulting redistribution of storage space has now increased Datastore1 from 71 GB to 185 GB free and has decreased the Datastore0 available (free) from 1 TB to 918 GB.
My question is how much FREE space do I have to leave on a datastore before it becomes an issue in terms of performance, etc. assuming that I don't need to migrate any additional vm attached storage space to it normally? So for Datastore0, e.g. which has 1 TB free right now, should I leave e.g. 20% of that space always free so 200 GB or can I use most of it with the understanding that Datastore0 at that point is basically not available for future space requirement considerations?
Datastore0 -> Capacity 1.80 TB, Free 1.00 TB
Datastore1 -> Capacity 1.80 TB, Free 71 GB
Datastore2 -> Capacity 1.99 TB, Free 105 GB
Datastore3 -> Capacity 1.99 TB, Free 1.55 TB
Note that I have space on two of the storage groups 0 and 3, but am down to 71 GB and 105 GB FREE on storage groups 1 and 2. I typically "migrate only the virtual machines to another host" for maintenance etc. but not their storage groups since the SAN does not go down for ESX updates and when the entire SAN needs to be rebooted then migrating storage would not help as we have only 1 SAN, i.e. I don't normally have to "migrate the virtual machines' storage to a compatible datastore or datastore cluster" nor have I had to "migrate the virtual machines to a specific host or cluster and their storage to a specific datastore or datastore cluster" but I just tried it and I was able to migrate a vm's storage only from Datastore1 (free 71 GB before migration) to Datastore0 (free 1 TB before migration) without issues and the resulting redistribution of storage space has now increased Datastore1 from 71 GB to 185 GB free and has decreased the Datastore0 available (free) from 1 TB to 918 GB.
My question is how much FREE space do I have to leave on a datastore before it becomes an issue in terms of performance, etc. assuming that I don't need to migrate any additional vm attached storage space to it normally? So for Datastore0, e.g. which has 1 TB free right now, should I leave e.g. 20% of that space always free so 200 GB or can I use most of it with the understanding that Datastore0 at that point is basically not available for future space requirement considerations?
ASKER CERTIFIED SOLUTION
membership
This solution is only available to members.
To access this solution, you must be a member of Experts Exchange.
ASKER
No problem and thanks for the insight plus great link... cheers...
Could you confirm if all the 4 Storage Groups have data? OR only 2 of them is using which are exhausted and the other 2 are free (1.8 each)?
You may not be able to redistribute free space from a storage group to another storage group, unless you break the storage group (raid group). If you can move data from one of the storage group to another storage group and delete\break that storage group to redistribute the disks to your already exhausted storage group.
You may need to identify each LUNs in a storage group, move the data from each LUNs to another LUN in a different Storage group. After making sure all the LUNs in that Storage group are clear, you can then delete the LUNs, then delete the Storage groups to free the disks. Then add the disks to the required (already exhausted) storage groups.
Hope EMC support will be able to assist you to achieve this.
Thanks,
Abhi