Hi,
I was looking into problem where we get duplicated output on RHEL5 Xen guest console logs after reboot and while testing my patch I found 2 more issues. I made a patch for problem 1, but I'm not sure what we can do about 2 and 3.
Do we still need rhts-compat? Is there an option to turn it off so people could try to run their tests without it and check if any tests need it?
Problem 1: Duplicate output in console logs on RHEL5 xen guests ---------------------------------------------------------------- http://gerrit.beaker-project.org/#/c/2630/
Problem 2: second serial port on guest is suffering from plymouth interference ------------------------------------------------------------------------------- See this Bug for details: Bug 1043261 - plymouth breaks serial console on kvm guest
And by the looks of it, it's problem with rhts-compat never completing: https://bugzilla.redhat.com/show_bug.cgi?id=1043261#c2
Problem 3: Duplicate output in console logs on RHEL6 ----------------------------------------------------- After this BZ: Bug 978419 - RFE: Change in rhel6 guest kickstarts
we have following in kickstarts: --append="console=ttyS0,115200 console=ttyS1,115200", which is problem, because we also pass "serial". Anaconda implicitly adds one more "console=ttyS0" if it finds "serial" on command line.
For example:
If you look at console.log, you can see that lines get duplicated: 2013-12-15 08:16:30,165 beah start_server: INFO ################################ 2013-12-15 08:16:30,165 beah start_server: INFO ################################ 2013-12-15 08:16:30,176 beah start_server: INFO # Starting a Controller... # 2013-12-15 08:16:30,176 beah start_server: INFO ################################ 2013-12-15 08:16:30,176 beah start_server: INFO Controller: BackendListener listening on :12432 2013-12-15 08:16:30,176 beah start_server: INFO # Starting a Controller... # 2013-12-15 08:16:30,176 beah start_server: INFO ################################ 2013-12-15 08:16:30,176 beah start_server: INFO Controller: BackendListener listening on :12432
Problem is again that plymouth keeps running and duplicates all output on console for each "console=ttyS0" it finds on kernel command line. Normally plymouth would be stopped when all runlevel 5 services start, but because rhts-compat runs forever then also plymouth runs forever: https://bugzilla.redhat.com/show_bug.cgi?id=1043261#c2
Regards, Jan
On 12/18/2013 06:19 PM, Jan Stancek wrote:
Hi,
I was looking into problem where we get duplicated output on RHEL5 Xen guest console logs after reboot and while testing my patch I found 2 more issues. I made a patch for problem 1, but I'm not sure what we can do about 2 and 3.
Thanks, we'll get an updated task published soon.
Do we still need rhts-compat? Is there an option to turn it off so people could try to run their tests without it and check if any tests need it?
Unfortunately, there's currently nothing as simple as a "no_rhts_compat" ks_meta var to forcibly skip turning it on. Adding one would probably be a good idea, though (we already have the conditional logic separated out to deal with RHEL7 and recent Fedora versions where it simply doesn't work because it is incompatible with systemd).
I suggest filing a bug for the console log issues, and we can at least add the new ks_meta option so people can workaround the problem, even if we can't figure out how to solve it automatically.
Regards, Nick.
beaker-devel@lists.fedorahosted.org