Server Fault Stack Exchange
Server Fault Stack Exchange

Questions tagged [perf]

11 questions
5
votes
1 answer

Better way to use Perf on a Program in Kubernetes/Docker?

I am running Kubernetes on CentOS 7, and it doesn't seem that the version of perf is namespace aware. If run perf on the PID on the host node (host node found with kubectl describe pods --namespace) I get an error about symbols not being found.…
Kyle Brandt
  • 83,619
  • 74
  • 305
  • 448
4
votes
1 answer

how to gracefully kill linux perf tool process

I'm trying to use the linux performance tool perf to monitor kvm events on a host where a guest VM exists. The command I invoke perf is perf kvm stat record After some time, I'll press Ctrl + C to stop it and it will show some info like how many…
Michael Tong
  • 141
  • 1
  • 2
2
votes
1 answer

High CPU usage by apache/httpd for light load, 'perf record' points at vmware/hardware

Running Apache 2.4 on RHEL7 kernel 3.10.0-1062, 4 CPU VMWare instance, doing very basic reverse proxying to WebLogic backend using WebLogic proxy plugin. Server is only pushing around 1 MByte/sec with a couple hundred users, listening for SSL and…
Dan
  • 21
  • 2
2
votes
0 answers

perf gives high copy_user_generic_string and clear_page_rep

I am building a server in c++ that reads a lot from disc. When recording with perf I get this report: 28.20% server [kernel.kallsyms] [k] copy_user_generic_string 18.14% server [kernel.kallsyms] [k] clear_page_rep …
joscul
  • 21
  • 2
1
vote
2 answers

3 different methods to count the number of context switches for a specific workload provide 2 different answers

I have a Centos7 system on a Google cloud machine, running a multithreaded database server that executes a particular set of queries in 70 secs (plus or minus 1 sec). I want to know how many context switches the server makes running this…
Richard Gostanian
  • 113
  • 1
  • 3
1
vote
0 answers

What's the best approach to find out in doing what a Linux OS process is stuck?

Oracle DBA here! No experience at all in troubleshooting Linux performance... I need to identify the bottleneck in an Oracle database running on Oracle Linux. The same procedure runs fine on other VMs, so the problem is specific to a specific VM…
Massimo Pasquini
  • 121
  • 1
  • 3
0
votes
2 answers

How to measure profit of using huge pages

I'm looking for a way of measuring performance impact/gain, in case of using huge pages. For example, I have 192GB RAM server, with 140GB allocated by huge pages. I run Postgres, with shared buffers in huge pages. Is there any way to check…
lesovsky
  • 243
  • 1
  • 7
0
votes
0 answers

High CPU of kworker/3:3+events (due to intel gpu drivers?)

I have a fresh install of ubuntu server 23.04 and I can see that CPU usage on idle is quite high (~0.5) after further investigation I saw that kworker proces periodically (something like every ~20sec kicks in and takes 100% of single core. After…
Bartoszek
  • 1
0
votes
0 answers

non-sudo perf sched always fails with "broken or missing trace data"

On my Ubuntu 22.04 (5.19.0-35-generic), it seems that perf sched is not able to parse its own perf.data output. Following the examples on Brendan Gregg's blog, recording works fine # perf sched record -- sleep 1 [ perf record: Woken up 1 times to…
Federico
  • 101
  • 2
0
votes
1 answer

Ping performance improves during iPerf3 network congestion test

I am beginning to learn about networking and testing network performance (specifically on containers). I am trying to learn how to perform good network tests to see networking capabilities of docker containers and whether they can support…
akastack
  • 3
  • 3
0
votes
0 answers

Multiple core overload on Ubuntu

I have a really strange behavior of the server. I run Apache, MySQL, WordPress on it. And recently, it has had a lot of freezes. MySQL/Apache shows high CPU usage (out of 24 cores - 2400% cpu% according to the htop). At first, I thought that the…
a b
  • 101
  • 1
  • 3