[DOCS-4534] Description and settings for readahead and I/O Scheduler for SSD devices Created: 16/Dec/14 Updated: 11/Jan/17 Resolved: 27/Jul/16 |
|
| Status: | Closed |
| Project: | Documentation |
| Component/s: | manual |
| Affects Version/s: | None |
| Fix Version/s: | 01112017-cleanup |
| Type: | Task | Priority: | Major - P3 |
| Reporter: | Eoin Brazil | Assignee: | Unassigned |
| Resolution: | Won't Fix | Votes: | 1 |
| Labels: | None | ||
| Remaining Estimate: | Not Specified | ||
| Time Spent: | Not Specified | ||
| Original Estimate: | Not Specified | ||
| Issue Links: |
|
||||||||
| Participants: | |||||||||
| Days since reply: | 8 years, 40 weeks, 2 days ago | ||||||||
| Description |
|
Hi All, Is it possible to update our production notes and related documentation to include a section on the best value for readaheads for an SSD (typically 0 but test as with 32 for HDD as depends on workload and doc sizes) and the use of the noop or deadline I/O scheduler based on testing with the use case? Thanks! |
| Comments |
| Comment by Emilio Scalise [ 11/May/15 ] |
|
I think that it could be useful also add an example script to change readahead on linux distributions that use init.d like this: I was trying to follow production notes on a centos 7 testing vm, and I lost a lot of time searching for a script that set readahead consistently between phyisical block devices and lvm and so on... And I think that many newbie customers has experienced this, seeing the number of CS tickets about this. Please note that I'm not complaining on the correct value of readahead to set (that need to be tested and determined by customers' tests), but for a quick way to set it for a wide range of linux distributions. |