Greg Howard here! I work here at Scale as a Technical Engineer. I have been with Scale for almost a year now. I work everything from technical support, services, and installs. I enjoy working here at Scale we have the best Managment, product and most of all customers!
Best posts made by ghoward
-
RE: Introductions
-
Support Tips: Scale HC3 Snapshot Scheduling and Rules
There are a few considerations to keep in mind when implementing a snapshot schedule for a VM. You will want to keep the overall available capacity on the cluster in mind when determining how many snapshots to retain. With the recurrences of the snapshot schedule in mind; one schedule can hold multiple recurrences from minutely to monthly configurations. As a best practice, separate different recurrences that may clash in different schedules.The shortest interval a snapshot recurrence that you should use is 5 minutes, anything less and it could cause unexpected behavior in the snapshot scheduling functionality. A VM can only have one snapshot schedule assigned to it at a time. You may have to create customized schedules for specific VMs depending on your backup needs.
The default snapshot schedule on 7.x is 15 minutes with a retention of 1 snapshot.This is designed to provide the lowest storage cost while still maintaining near real-time DR capabilities. It is important to evaluate multiple factors when determining your snapshot schedules such as Storage availability (source and target cluster), your current DR requirements, and network bandwidth limitations (when utilizing replication). It is also important to remember that different snapshot schedules may benefit VMs with different DR requirements.
-
RE: Support Tips/Tricks and maybe a Treat or two!
Greg here from the support team! We created this space for the Scale Legion community members to ask questions, give comments and also for us to put out tips or tricks that you might benefit from. We will be putting out information weekly and check this daily. I would like to stress that if you are having an issue with you Scale cluster that you do call Scale support or send in a case.
-
Support Tips: Network Performance Troubleshooting
Troubleshooting networking performance issues can be tricky. It is always recommended to engage Scale Computing support if you believe there is is a networking performance issue. There are a few key items to consider:
-
One VM being affected or all VMs?
-
One node being affected or all nodes?
-
Are there any errors in the switch logs?
-
Are there performance issues elsewhere on the network?
-
What color are the NIC lights?
Having a clear picture of the issue you are experiencing before engaging support can be very helpful for resolving the issue quickly. If you do have access to the HC3 cluster verify that the node or cluster is physically on. Verify that all cables are connected to the nodes as well as the switch and fully seated in the ports. If everything appears to be connected you will want to check the Network Interface Card (NIC) A solid green light indicates connectivity is present and a flashing amber lights indicate collisions. Check the switch logs to see if there are any network connectivity issues or the switch is dropping the connection to the cluster.
-
-
Support Tips: Access Port vs Trunk Port
Access Port
-
Allows one untagged VLAN to pass through it
-
Access ports are typically used for “edge” devices such as computers because of their lack of ability to handle tagged traffic
-
Any tagged traffic sent to or from the interface will be dropped
-
Can be used to isolate a network segment such as a DMZ
Trunk Port
-
Referring to a VLAN trunk rather than link aggregation
-
A trunk port allows one untagged and all tagged VLANs to pass through it as long as they are added to the port configuration
-
Trunk ports are typically used for connections between switches to allow for multiple VLANs to communicate but can be used for any device that can manage multiple VLANs and include a VLAN tag on the frames
-
The untagged traffic will be sent on the native VLAN
-
-
Support Tips: Gold Master Images for Scale HC3
Hello all, This week we are going to cover "golden master images". A brief description of a golden image it's a template for a virtual machine (VM). A golden image may also be referred to as a clone image, master image or base image.
Given the nature of VSDs and thin provisioning and snapshotting it is recommended to utilize a “Gold Master” image or clone when creating VMs. This is also beneficial when utilizing VMs as targets for migrations using HC3 Move. The Gold Master should contain any OS updates and critical applications that will be shared among all the machines. It is also recommended to utilize a sysprep to change the local SID to prevent issues down the road. There is a knowledge document that outlines this process called “Using Sysprep to Generate a new SID”.
There are also some virtualization best practices for VMs that would be beneficial in most instances when creating a Gold MasterBest practices for performance for Gold Master Image list item:
-
Disable Fast Startup (if applicable to Windows version):
- Fast Startup is a feature introduced with Windows 8 that can improve boot times on physical PCs
- However, as a VM, this can cause problems when initializing hardware after shut down (this can especially be true if you are assigning physical devices to virtual machines) -
Disable Hiberfil.sys:
- Hibernation suspends the Windows OS to RAM so that on resume, the entire system state returns as it was
- Hiberfil.sys will be created on the root of your C:\ drive that can take up a massive amount of space and generate lots of disk IO activity for a feature that likely never be leveraged -
Disable Windows Indexing:
- The Windows indexing service (aka Windows Search) can cause unnecessary disk IO inside your VM
- To improve overall VM performance, it is recommended to disable this service from running in Windows if it is not planned to be used -
Disable Automatic Disk Defragmentation:
- If Windows disk defragmenter is set to automatically run on a schedule, you should disable it, this serves no function in a virtual environment -
Enable High Performance Power Mode:
- In a virtual environment there is little to no benefit in leveraging power management features
-
Latest posts made by ghoward
-
Support Tips: Access Port vs Trunk Port
Access Port
-
Allows one untagged VLAN to pass through it
-
Access ports are typically used for “edge” devices such as computers because of their lack of ability to handle tagged traffic
-
Any tagged traffic sent to or from the interface will be dropped
-
Can be used to isolate a network segment such as a DMZ
Trunk Port
-
Referring to a VLAN trunk rather than link aggregation
-
A trunk port allows one untagged and all tagged VLANs to pass through it as long as they are added to the port configuration
-
Trunk ports are typically used for connections between switches to allow for multiple VLANs to communicate but can be used for any device that can manage multiple VLANs and include a VLAN tag on the frames
-
The untagged traffic will be sent on the native VLAN
-
-
Support Tips: Network Performance Troubleshooting
Troubleshooting networking performance issues can be tricky. It is always recommended to engage Scale Computing support if you believe there is is a networking performance issue. There are a few key items to consider:
-
One VM being affected or all VMs?
-
One node being affected or all nodes?
-
Are there any errors in the switch logs?
-
Are there performance issues elsewhere on the network?
-
What color are the NIC lights?
Having a clear picture of the issue you are experiencing before engaging support can be very helpful for resolving the issue quickly. If you do have access to the HC3 cluster verify that the node or cluster is physically on. Verify that all cables are connected to the nodes as well as the switch and fully seated in the ports. If everything appears to be connected you will want to check the Network Interface Card (NIC) A solid green light indicates connectivity is present and a flashing amber lights indicate collisions. Check the switch logs to see if there are any network connectivity issues or the switch is dropping the connection to the cluster.
-
-
Support Tips: Gold Master Images for Scale HC3
Hello all, This week we are going to cover "golden master images". A brief description of a golden image it's a template for a virtual machine (VM). A golden image may also be referred to as a clone image, master image or base image.
Given the nature of VSDs and thin provisioning and snapshotting it is recommended to utilize a “Gold Master” image or clone when creating VMs. This is also beneficial when utilizing VMs as targets for migrations using HC3 Move. The Gold Master should contain any OS updates and critical applications that will be shared among all the machines. It is also recommended to utilize a sysprep to change the local SID to prevent issues down the road. There is a knowledge document that outlines this process called “Using Sysprep to Generate a new SID”.
There are also some virtualization best practices for VMs that would be beneficial in most instances when creating a Gold MasterBest practices for performance for Gold Master Image list item:
-
Disable Fast Startup (if applicable to Windows version):
- Fast Startup is a feature introduced with Windows 8 that can improve boot times on physical PCs
- However, as a VM, this can cause problems when initializing hardware after shut down (this can especially be true if you are assigning physical devices to virtual machines) -
Disable Hiberfil.sys:
- Hibernation suspends the Windows OS to RAM so that on resume, the entire system state returns as it was
- Hiberfil.sys will be created on the root of your C:\ drive that can take up a massive amount of space and generate lots of disk IO activity for a feature that likely never be leveraged -
Disable Windows Indexing:
- The Windows indexing service (aka Windows Search) can cause unnecessary disk IO inside your VM
- To improve overall VM performance, it is recommended to disable this service from running in Windows if it is not planned to be used -
Disable Automatic Disk Defragmentation:
- If Windows disk defragmenter is set to automatically run on a schedule, you should disable it, this serves no function in a virtual environment -
Enable High Performance Power Mode:
- In a virtual environment there is little to no benefit in leveraging power management features
-
-
Support Tips: Scale HC3 Snapshot Scheduling and Rules
There are a few considerations to keep in mind when implementing a snapshot schedule for a VM. You will want to keep the overall available capacity on the cluster in mind when determining how many snapshots to retain. With the recurrences of the snapshot schedule in mind; one schedule can hold multiple recurrences from minutely to monthly configurations. As a best practice, separate different recurrences that may clash in different schedules.The shortest interval a snapshot recurrence that you should use is 5 minutes, anything less and it could cause unexpected behavior in the snapshot scheduling functionality. A VM can only have one snapshot schedule assigned to it at a time. You may have to create customized schedules for specific VMs depending on your backup needs.
The default snapshot schedule on 7.x is 15 minutes with a retention of 1 snapshot.This is designed to provide the lowest storage cost while still maintaining near real-time DR capabilities. It is important to evaluate multiple factors when determining your snapshot schedules such as Storage availability (source and target cluster), your current DR requirements, and network bandwidth limitations (when utilizing replication). It is also important to remember that different snapshot schedules may benefit VMs with different DR requirements.
-
RE: Support Tips/Tricks and maybe a Treat or two!
Greg here from the support team! We created this space for the Scale Legion community members to ask questions, give comments and also for us to put out tips or tricks that you might benefit from. We will be putting out information weekly and check this daily. I would like to stress that if you are having an issue with you Scale cluster that you do call Scale support or send in a case.
-
RE: Introductions
Greg Howard here! I work here at Scale as a Technical Engineer. I have been with Scale for almost a year now. I work everything from technical support, services, and installs. I enjoy working here at Scale we have the best Managment, product and most of all customers!