The URL referenced in the lab for the 'stress' tool is broken.
yes indeed, whether his link is dead forever or temporarily I don't know, but you can always use the better stress-ng, which you can get from https://kernel.ubuntu.com/git/cking/stress-ng.git/ . AFAIK, it is completely backwards compatible.
Maybe he left Harvard
Couldn't handle the stress?
Hi @troy_dowler ,
You also could find "stress" in the repo of your distro. I'm on Ubuntu 19.04 and I see here:
$ apt-cache show stress-ng
Maintainer: Colin King [email protected]
Depends: libaio1 (>= 0.3.93), libapparmor1 (>= 2.10), libbsd0 (>= 0.5.0), libc6 (>= 2.29), libipsec-mb0, libsctp1 (>= 1.0.10+dfsg), zlib1g (>= 1:1.1.4)
I don't know why, but if you click on the link above ( https://kernel.ubuntu.com/git/cking/stress-ng.git ) it seems to fail, but if you copy and paste into your browser it works. Beats me
Might have to use the the method raised by @Coop
yum search stress-ng
Loaded plugins: fastestmirror Loading mirror speeds from cached
hostfile ** base: centos.melbourneitmirror.net ** extras: centos.melbourneitmirror.net ** updates: centos.melbourneitmirror.net Warning: No matches found for: stress-ng No matches found
I checked on pkgs.org and it looks like pre-made packages are available for most of the common distrobutions with the exception of CentOS8. This list does not mention RHEL8.
Troy, would you happen to be using CentOS8?
Hi Lee, I'm using CentOS7. Looks like I will need to a. compile the package, or b. install the stress-ng package and the dependencies. Not really sure right now how to go from here. I found this url which is what I'm trying to work off.
Both stress and stress-ng are in the EPEL repository for CentOS7. so a simple yum install would work if you turn on the epel repo. You can download the epel-7 repo a bunch of places or do the one line command:
sudo yum install https://dl.fedoraproject.org/pub/epel/epel-release-latest-7.noarch.rpm
@Coop, thanks for that. I have now stress-ng installed and hopefully i can complete the lab now.
Confession time. I did try various things to make stress-ng work including what Coop mentioned Oct 03, but never got it working and gave up. Now, I realised that Lab 13.1 requires 'stress-ng' and not 'stress', so we have inconstancy with the Labs. Anyhow hopefully at least posting this matter may help somebody in some shape or form if they also run into this. I also forgot to mention that I'm using Raspberry PI 4 as my linux-box and up until lab 11.1 it has been fine.
Unless there is something weird on the RPI, this lab does not require stress-ng. It was developed originally on stress and has been working on that platform for quite a few years. The thing that has changed is the dififculty of obtaining stress as perhaps its author is moving on and letting people use stress-ng which is essentially a superset of stress
Well, that explains why you have been having some trouble with the tools... Just wondering, what OS are you running on the Raspberry PI 4? Anyway, that may not be the best device to make the labs, as the OS's for these type of devices uses to be limited versions when talking about software.
I'm using CentOS 7 on the Raspberry PI. Error below properly explains the issue. Do you know any other similar tools i can use to artificially load the system?
[[email protected] ~]$ yum deplist stress-ng
Loaded plugins: fastestmirror, verify
Loading mirror speeds from cached hostfile
epel/armhfp/metalink | 54 kB 00:00:00
Could not parse metalink https://mirrors.fedoraproject.org/metalink?repo=epel-7&arch=armhfp&infra=stock&content=altarch error was
No repomd file
File /var/tmp/yum-troy-R5bUFz/armhfp/7/epel/metalink.xml does not exist
I'm not an expert on this, but epel does not officially exist for arhfp, see [https://wiki.centos.org/SpecialInterestGroup/AltArch/armhfp#head-b68dc016c9892844d739cbf121d360b0e3371c4f]
Basically you are using an unsupported platform for this course which is intended to be on an enterprise level distribution, so while it is fun, we have not endeavored to figure out these vagaries. So good luck and enjoy, but unless we have a RPI enthusiast paying attention there's a limit to how much we can debug this stuff for you. At any rate, compiling from source is a better solution as the code itself for stress and stress-ng is not very architecture dependent anyway
I strongly suggest you to use a vm for doing the labs. We don't know what else is gonna fail in your device, and it's not supported as well.
*But, if you want to try, you can get the stress-ng sources from the git repo here:
I can't assure it will work anyway, and I still suggest to switch to a vm.
I haven't used other tool to stress the system cpu, but there are a lot to stress a web server and services in the could, but it's not the case.