[SERVER-6331] mongoperf can't saturate drive with writes Created: 06/Jul/12  Updated: 15/Mar/16  Resolved: 15/Mar/16

Status: Closed
Project: Core Server
Component/s: Tools
Affects Version/s: 2.1.0
Fix Version/s: None

Type: Bug Priority: Minor - P4
Reporter: Jared Rosoff Assignee: Unassigned
Resolution: Won't Fix Votes: 1
Labels: None
Remaining Estimate: Not Specified
Time Spent: Not Specified
Original Estimate: Not Specified
Environment:

Amazon Linux


Attachments: PNG File mongoperf.png    
Operating System: Linux
Participants:

 Description   

When running some tests on EC2, I was using mongoperf to evaluate disk performance. mongoperf was run with settings

{nThreads:16, w:true, fileSizeMB:10000}

. Despite ramping up the number of threads (tried up to 64 threads) mongoperf could never do more than ~1100 ops / second. However, if I run multiple mongoperf processes, I was able to get the aggregate up to 5000 ops / second. Seems like 1 mongoperf instance should be able to saturate the disk. I've attached a chart showing iops reported from iostat during the test. the "steps" on the left hand side show the aggregate throughput as i went from 3 -> 4 -> 5 mongo perf instances. the "valleys" are the time during which the new mongoperf is allocating the data file. the bulk of the graph is the performance with 5 mongo perf's running simultaneously.



 Comments   
Comment by Ian Whalen (Inactive) [ 15/Mar/16 ]

Apologies for the delay in responding on this issue, but we've decided not to proceed with any further enhancements to mongoperf in its current form - as such we're closing this as Won't Fix. We aren't moving the existing code or removing the binary, but any changes going forward would almost certainly involve first moving the existing functionality to a newly written tool. We made a similar decision with our other tools in 3.0.0 and have been very happy with the results.

We don't have a specific timeline for this work, but you can vote on and add yourself as a watcher to TOOLS-1103 to be informed of updates as we weigh its priority against other projects that our Tools team are working on.

Comment by Brennan Pang [ 24/Jan/14 ]

I found the same issue on a CentOS 6.3 physical host, even running 1 mongoperf instance with a ridiculous number of threads (15k) I could only push 22MB/sec. Once I spun up multi instances of mongoperf with separate large datafiles, I could push past 400MB/sec on our SSD array.

Generated at Thu Feb 08 03:11:22 UTC 2024 using Jira 9.7.1#970001-sha1:2222b88b221c4928ef0de3161136cc90c8356a66.