-
-
Notifications
You must be signed in to change notification settings - Fork 101
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
pbTests: Use 'homemade' Solaris box #2405
Conversation
As an update, as I found whilst looking at #2411 , the new box does have |
FYI, the latest commit will increase the timeout for the vagrant box. By default the timeout is 5 minutes, however apparently it is taking a bit longer on the VPC machine. If it still doesn't boot after 10 minutes, something is probably very wrong. Latest run: VPC-1352 |
Okay, I've managed to get a JDK build in VPC-1355 . This managed to successfully build JDK8/HS ; however, the tests weren't able to run, I believe due to the problem that is mentioned in #2210 . I also noticed that in the actual tests, were
Also, I'm not 100% certain the long boot times fo rthe VM are fixed. I was unable to recreate the issue on my machine, nor directly using the VPC machine, and it seems the boot time varies wildly, wherein it took >10 minutes to boot in VPC-1353, and 3 minutes to boot in VPC-1355 |
Testing running the tests, locally, I've determined a few things: we need to install
Googling has told me that this means In addition, some kind of perl
On something like RHEL, it'd be For reference, the perl version that's being used is EDIT: |
Carrying on from that, it seems we need to install a newer version of ant:
The test scripts do carry on, beyond that point, however it skips the test:
|
Hmmm I would have thought we'd be pulling ant 1.10.5 into /usr/local/ like we do elsewhere ... Unless we've got two versions installed and it's picking up the wrong one by default - which seems the most likely option and would explain why it didn't find ant-contrib properly too. |
Ah yes, I've found it; you were right. Installing it via PKGUtil just puts it in |
Testing that locally, (i.e. adding to the path
AFAIK, no special options have been put on the JVM, so I guess the test is skipping due to the platform. There doesn't appear to be any Solaris / SunOS mentioned in the |
Ref: The latest committ. As said by @sophia-guo , Solaris can't run the Theoretically, with this, the VPC Solaris run should now work - VPC Run: VPC-1632. |
It worked (!!!). Ready for review. |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Nothing too controversial in here - just the result of a lot of debugging to figure out why things were failing :-)
I'd suggest we want some docs on the process of setting up the Vagrant machine somewhere too, but other than one quoting suggestion I'm ok for this to go in :-)
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Let's go back to a simpler test instead of the whole jdk_math
suite
a84ba1d
to
712ae49
Compare
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
LGTM
Ref: https://adoptium.slack.com/archives/C53GHCXL4/p1637309425055700
The Solaris box we were using didn't have any of the X11 libraries that is required to build a JDK. Therefore, with the help of @sxa and @gdams , I've made a new vagrant box; which I have installed to every
vagrant*
user on the VPC machine. This box has a lot more libraries located in/usr/openwin/
, so, fingers-crossed it means we can actually build a JDK on there!Creating this box has also allowed for it to be setup how we want it. Therefore, the compiler is pre-installed on the box so we don't have to do that whenever creating the VM, we have a newer version of SSH on there, so we don't have to change the ssh
KexAlgorithm
, and the VBox Guest Additions are up to date.The
.box
file that was used to add the box to the machine, is located on theinfra-ibmcloud-vagrant-x64-1
machine, at/home/will/solaris10_homemade.box
, in case any additional config needs to be put in. The PR is in draft as I want to get some working JDK builds before merging it :-)Please note; I've taken the
adoptopenjdkSol10.vm.hostname
argument out of the Vagrantfile, as this was causing the/etc/hostname
file to be deleted. Deleting this file causes the network interface configuration command to fail, due tosudo
'sunable to resolve host
issue. The hostname that's been set on the box isadoptSolaris10
, and I haven't seen any additional errors come up from removing the argument.