I am trying to set the RAID virtual disk settings, and in this case cannot determine which is more optimized;
Read Policy: Out of Read-Ahead, No-Read-Ahead and Adaptive Read-Ahead, the default is Read-Ahead.
I will do a lot of sequential writes initially, writing for the virtual machine hard drive A blank image (for example, assuming /dev/0 is 30GB), so Read-Ahead seems to be good at first. But in a virtual machine, reading anywhere in the file system can be randomly read because they are IIS and MSSQL Server, so maybe No-Read-Ahead is a better idea?
Now I think Adaptive Read-Ahead will be better than compromise, but I don’t know much about this option. How does it compare to other options in terms of performance?
Write Policy: write-back caching, write-through caching, the default is write-back caching.
The default value of write-back caching It’s more secure than write-through cache, but the performance cost is higher. My thinking is that, for example, in the event of a power outage, it seems more likely to be in my mind (that’s why I need some clarification!) to enable The guest virtual machine in the write-back cache is damaged, so should I support write-through?
I have searched around and obviously there is no clear answer, so I want to find the method that best suits my situation.
Secondly, if there is any possibility, you always want write-back cache. It will confirm the write immediately after saving to the cache, and then add it as soon as possible They are forwarded to disk. However, if you lose the power to the cache, it is not safe. If you want consistent data when you come back from a power failure, battery backup cache is a necessary condition for write-back caching.
I have some Dell PowerEdge 2950s that contain 4x 15k, 150GB Cheetah SAS drives. They will be VM hosts, CentOS running ESXi and Windows Server 2k8 clients. Some guests will host IIS servers , And other MSSQL servers.
I am trying to set the RAID virtual disk settings, and in this case cannot determine which is more optimized;
Read Policy: Out of Read-Ahead, No-Read-Ahead and Adaptive Read-Ahead, the default is Read-Ahead.
I will do a lot of sequential writes initially, writing for the virtual machine hard drive A blank image (for example, assuming /dev/0 is 30GB), so Read-Ahead seems to be good at first. But in a virtual machine, reading anywhere in the file system can be randomly read because they are IIS and MSSQL Server, so maybe No-Read-Ahead is a better idea?
Now I think Adaptive Read-Ahead will be better than compromise, but I don’t know much about this option. How does it compare to other options in terms of performance?
Write Policy: write-back caching, write-through caching, the default is write-back caching.
The default value of write-back caching It’s more secure than write-through cache, but the performance cost is higher. My thinking is that, for example, in the event of a power outage, it seems more likely to be in my mind (that’s why I need some clarification!) to enable The guest virtual machine in the write-back cache is damaged, so should I support write-through?
I have searched around and obviously there is no clear answer, so I want to find out the method that best suits my situation.
First of all, Hypervisors are sometimes called “IO mixers”. They make it difficult for any algorithm to predict what they will do next, because they will not let each VM have its own real SCSI disk. Therefore, you definitely do not want to directly Read. I don’t know what your system thinks adaptive pre-reading is, but if it is a setting that will detect when the sequential read is completed and read a little bit conservatively, it may be fine.
Secondly, if there is any possibility, you always want write-back caching. It will confirm the writes immediately after saving to the cache, and then forward them to disk as soon as possible. However, if the cache is lost Power supply, this is not safe. If you want to get consistent data when you come back from a power failure, the backup battery cache is a necessary condition for write-back cache.