This session provides an opportunity to learn and ask questions about Cisco Unified Computing System in general and the blades in more specific. Sivakumar will cover questions on design, configuration and troubleshooting.
The Cisco Unified Computing System™ is a next-generation data center platform that unites compute, network, storage access, and virtualization into a cohesive system designed to reduce total cost of ownership (TCO) and increase business agility. The system is an integrated, scalable, multichassis platform in which all resources participate in a unified management domain. Managed as a single system whether it has one server or hundreds of servers with thousands of virtual machines, the Cisco Unified Computing System decouples scale from complexity.
Ask questions from Monday, November 30 to Friday December 11, 2015
Featured Experts
Sivakumar Sukumar is an experienced Customer support engineer in High-Touch Technical Services (HTTS) team supporting Data center Products. His areas of expertise include Cisco UCS, UCS central, Cisco Nexus 9000, Nexus 5000, Nexus 3000, Nexus 2000, MDS SAN switches, Cisco Application Control Engine (ACE), Cisco Wide Area Application Services (WAAS), Cisco Content Switching Module, Cisco Content Services Switches, and other content products. He has been with Cisco for more than 6 years, working with major customers to help resolve their issues related to content products. Prior to Cisco, he worked with IBM and other service providers managing their networks. Sivakumar holds a bachelors in Information Technologies from SSNCE Anna University Chennai India, and these Cisco certifications: CCIE (#44288) (in Datacenter), CCNA, CCNP, RHCE and VCP.
Sivakumar might not be able to answer each question due to the volume expected during this event. Remember that you can continue the conversation on the Data Center sub-community
Find other https://supportforums.cisco.com/expert-corner/events.
** Ratings Encourage Participation! **
Please be sure to rate the Answers to Questions
Hello Sebastian,
Good question!
Please refer install and upgrade guide for information for running UC apps on UCS
To achieve good performance please make sure the drivers are updated as per HCL
Find ESXi Version and RAID/NIC/HBA Driver
Some servers may come with a RAID array configured from the factory. Even if this is the case you should verify that the array settings match those specified in the Cisco Collaboration on virtualized servers document.
Here are some very useful tips to achieve good raid performance on LSI RAID controllers on UCS
What can cause a bad raid performance?
One of the most common issues for a bad system performance is having the cache being disabled!
" A cache is a temporary memory (storage area) where frequently accessed data can be stored for rapid acces. "
The first thing to check is if the RAID Cache is ON, and enable it if it is not.
* Some RAID controllers without a BATTERY BACKUP UNIT on board switch the cache OFF by default!
* Some RAID controllers need to set more than just one single parameter in order to enable the cache.
RAID controllers with LSI chipset have following options:
Default Write Policy: Write Through; Write Back
Cache Policy: Cached; Direct
Disk cache policy: Enabled; Disabled
Read Policy: Readahead; None; Adaptive
In order to get the cache enabled, please go to RAID Controller BIOS and make sure following options are selected:
Default Write Policy: Write Back
Cache Policy: Cached
Disk cache policy: Enabled
Read Policy: Adaptive / In some cases Readahead can be better.
Example:
Configure Sequential Access Patterns
> Use sequential access patterns in the range of 64 KB to 1MB
> Use random access patterns in the range of 0.5 KB to 8 KB
More Drives > To achieve maximum performance, connect enough drives to saturate the MegaRAID controller.
Use Adaptive Read Policy > This policy should be used for alll configurations and directs the controller to use read-ahead if the two most recent disk access ocurred in sequential sectors. If all read requets are random, the algorithm automatically changes to No Read Ahead while still evaluating all requests for possible sequential operation. LSI does not recommend using the Always Read Ahead policy.
Write Policy: Write-Through > With the Writte-through caching strategy, data is written to disk before a completion status is returned to the host operating system. This is considered more secure, because a power failure is less likely to cause undetected drive write data loss. Recommended option Write-through policy for RAID 0, RAID 1 and RAID 10 configurations with streaming or sequential data writes, because it can avoid copying data into a cache.
Write Policy: Write-Back > With the Write-Back caching strategy, a completion status is sent to the host operating system as soon as data is written to the RAID cache. DATA is written to the disk when it is forced out of the controller cache memory. Write-back is more efficient in environments with "bursty" write activity.
Using the Write-back policy to achieve optimum performance in RAID 0, RAID 1, and RAID 10 configurations with transactional (random real world) reads. Recommended using the Write-back policy to achieve optimum performance in all RAID 5 and RAID 6 configurations because it can improve the performance of data redundancy generation.
Cache Policy: Direct I/O recommended for all RAID level configurations.
HTH
Siva
