Description of problem: When I try to install rhel-8 manully by specify: Kickstart Metadata: manual Kernel Options: vnc I always hit this error: --- An error occurred during reading the kickstart file: One of --drives or --only-use must be specified for ignoredisk command. Pane is dead [anaconda]1:main* 2:shell 3:log 4:storage-log >Switch tab: Alt+Tab | Help: F1 --- By checking the ks file, I find this command: ignoredisk --interactive The --interactive option has ben deprecated since version Fedora29 (from office doc), so I doubt RHEL-8 has deprecated this option too. If so, we generate a wrong ks file for rhel-8, which cause we can't install rhel-8 interactively. Version-Release number of selected component (if applicable): RHEL-8.0-20180822.n.1 How reproducible: 100% Steps to Reproduce: 1. Set Kickstart Metadata to manual 2. Set Kernel Options to vnc Actual results: Can't install rhel-8 Expected results: install rhel-8 successfully Additional info: I don't know if I report this bug on the right component, feel free to correct me.
Oh, nice find! Thanks for reporting this. The next question will be, why did we have it originally, and what does Anaconda want us to use instead...
Is this blocking any of your work? I'm asking because you set the severity to high.
(In reply to Roman Joost from comment #2) > Is this blocking any of your work? I'm asking because you set the severity > to high. We have a special machine with 500T storage, it's a machine we used to do manual test. I've installed RHEL-6 and RHEL-7 systems on it, and have set up the test env on different systems. Now I want to install RHEL-8 on it too, and don't want to destroy other systems. So I have to turn to 'Kickstart Metadata: manual', then manually install through vnc. I think it will affect all similar requirement, just because most of QEs are working on RHEL-7, so they haven't hit this issue. But I think maybe there're other ways to intall system bypass the beaker kickstart file. If there's one, this's not a blocker for me. Thanks, Zorro
(In reply to Dan Callaghan from comment #1) > Oh, nice find! Thanks for reporting this. > > The next question will be, why did we have it originally, and what does > Anaconda want us to use instead... The doc which I read is (I didn't find rhel-8 related doc): https://pykickstart.readthedocs.io/en/latest/kickstart-docs.html#ignoredisk ---- ignoredisk [--drives IGNOREDISK] [--only-use ONLYUSE] [--interactive] ... ... --interactive Allow the user manually navigate the advanced storage screen. New in version RedHatEnterpriseLinux6. Deprecated since version Fedora29. ---- Acturally I don't learn about rhel-8 kickstart, I just suppose it's same with Fedora 29 now, due to I hit this error. How about don't generate any ignoredisk command line, if the user didn't specify --drives or --only-use? Maybe we can give it a try on RHEL-6/7 and RHEL-8. That's just my guess. Feel free to talk with anaconda team and correct me. Thanks, Zorro
You could explore if you could pass an empty string to --ignoredisk: https://beaker-project.org/docs/user-guide/customizing-installation.html#kickstart-metadata (you have to scroll down) like so: ks_meta="ignoredisk=''" Alternatively, you could replace the entire kickstart file with: https://beaker-project.org/docs/user-guide/customizing-installation.html#custom-kickstart-templates which comes with the down side, that you need to get your kickstart file right, otherwise it could fail in other weird ways.
(In reply to Roman Joost from comment #5) > You could explore if you could pass an empty string to --ignoredisk: > > https://beaker-project.org/docs/user-guide/customizing-installation. > html#kickstart-metadata (you have to scroll down) > > like so: > > ks_meta="ignoredisk=''" Thanks, I'll give it a try. > > Alternatively, you could replace the entire kickstart file with: > > https://beaker-project.org/docs/user-guide/customizing-installation. > html#custom-kickstart-templates > > which comes with the down side, that you need to get your kickstart file > right, otherwise it could fail in other weird ways. Thanks, maybe I misunderstand. I think the things between <ks_append>...</ks_append> will append into original ks file, it won't replace it. So if the original file has problems, it won't be fixed by this method. Is there a way to replace the ks file which beaker uses by default? Thanks, Zorro
> > Alternatively, you could replace the entire kickstart file with: > > > > https://beaker-project.org/docs/user-guide/customizing-installation. > > html#custom-kickstart-templates > > > > which comes with the down side, that you need to get your kickstart file > > right, otherwise it could fail in other weird ways. > > Thanks, maybe I misunderstand. I think the things between > <ks_append>...</ks_append> will append into original ks file, it won't > replace it. This is correct. > So if the original file has problems, it won't be fixed by this > method. Is there a way to replace the ks file which beaker uses by default? Yes by my link mentioned above. You can replace the entire kickstart file with your own if you like.
(In reply to Roman Joost from comment #7) > > > Alternatively, you could replace the entire kickstart file with: > > > > > > https://beaker-project.org/docs/user-guide/customizing-installation. > > > html#custom-kickstart-templates > > > > > > which comes with the down side, that you need to get your kickstart file > > > right, otherwise it could fail in other weird ways. > > > > Thanks, maybe I misunderstand. I think the things between > > <ks_append>...</ks_append> will append into original ks file, it won't > > replace it. > > This is correct. > > > So if the original file has problems, it won't be fixed by this > > method. Is there a way to replace the ks file which beaker uses by default? > > Yes by my link mentioned above. You can replace the entire kickstart file > with your own if you like. Oh, the 'ks=<url>'. I never used it before, I'll try, thanks! So this bug is not a blocker for me. Reduce the bug Severity if there's not more concern from others. But this's still a bug, hope it can be fixed. Thanks, Zorro
*** This bug has been marked as a duplicate of bug 1638258 ***