On Mon, Jan 28, 2013 at 7:37 PM, Michael Hudson-Doyle <[email protected]> wrote: > Paul Sokolovsky <[email protected]> writes: > >> I tried to investigate how it happened that lava-dispatcher doesn't try >> to set hostname, and actually found it used to, but that was tied to >> recognition of test prompts, and once that was refactored, setting >> hostname also gone, see >> http://bazaar.launchpad.net/~linaro-validation/lava-dispatcher/trunk/revision/402 >> (grep for "_customize_ubuntu" - it got factored out to another file, >> and lost /etc/hostname setting in progress). >> >> So, please let me know if the above makes sense, and should be files as >> lava-dispatcher bug, or are there other reasons for not setting a >> hostname. > > Seems reasonable. Do you want hostname set to the device name > ("panda01") or just something generic? Previously the dispatcher just > set it to "linaro" I think...
I don't feel we should start encoding anything useful into the hostname that doesn't help LAVA to reliably match the prompt. All other ideas that come to our mind that read "use the prompt to identify..." probably just indicate the need for something that gives meta info about the id/host at hand. Here, idea would be to make a (busybox) tool we install on all targets that dumps information about the host ... like ec2metainfo or lsb_release -r ... I assume we already talked about that in the past? > > Cheers, > mwh -- Alexander Sack Director, Linaro Platform Engineering http://www.linaro.org | Open source software for ARM SoCs http://twitter.com/#!/linaroorg - http://www.linaro.org/linaro-blog _______________________________________________ linaro-validation mailing list [email protected] http://lists.linaro.org/mailman/listinfo/linaro-validation
