I have an old core2duo machine around here which doesn't seem to accept omni installation media. I dd'd the installation image to a flash drive, but the machine doesn't take it. I can boot other machines from it though. Do I have any other option that network booting?

you could try to find out why it fails and fix that

i guess a core2duo is pre-uefi, so i'd check that the installation image supports that. and i'd check whether the installation image and/or the flash drive work correctly otherwise, like on another system

there's a uefi toggle in bios, but it doesn't seem to have any impact and it seems to work on (at least one) other machine

in that case it would be interesting to know how it fails. what error message do you get?

none really, it just skips directly to booting from network

the bootorder is usb, then disks, then network

is there a bios boot menu that you can use to choose one particular device, and that device only?

not exactly. I tried simulating something like this by unplugging everything else, but no luck. Then it goes straight to "no boot device found, insert boot disc and press enter to reboot" which is a bios thing I assume

If you haven't already tried, write the .iso image to the USB drive. If that's the one you used, try the .usb-dd one

have you tried booting that machine with a different flash drive, or the same flash drive but with a different image (of a different OS, perhaps)?

andyf: tried that too, no luck there either

I tried the same machine with a different flash drive and a different image (some linux) and that worked, but yeah, worth a try I guess

hi, I installed nginx a while back, and noticed the smf method file doesn't actually take the config file into account? so I modified it to do so, and I'm interested in maybe getting this modification in the package

and I'd also like to hook up `svcadm refresh` to dnsmasq to send SIGTERM to the process, since this is how dnsmasq reloads it's config file

Hi patrikr. We're definitely interested in fixes and enhancements like that.

is the github repo the upstream or is there a different repo? i assume i just create a pull request

github.com/omniosorg/omnios-extra is where the packages are built from.

thanks, found the correct file I think

I'm sure nginx is supposed to use the config file property, to make it possible to run multiple instances of the service.

Maybe nobody has actually used that

yeah, I tried changing the smf prop but it didn't actually do anything

my fix is just exec ${NGINX} -c ${CONF_FILE} 2>&1 , which I suppose doesn't handle an empty config/file but not sure if that should be allowed

oh, just saw the script actually checks if it is set or not so shouldn't matter

It probably also needs "-g pid $PIDFILE;" adding, if PIDFILE is set

It probably also needs "-g 'pid $PIDFILE;'" adding, if PIDFILE is set

Well, some variation

It obviously depends on whether we want to think of PIDFILE as telling it which file to clean up, or as a configuration property too.

that could be why I've been having issues where I could've sworn I stopped nginx but it seems to still be running

since the PIDFILE seems to be set by default, is adding '[ ! -f "$PIDFILE" ] && exit $SMF_EXIT_ERR_CONFIG' right under the conf_file check reasonable?

I don't think that's quite right because when the service starts up there might not be a pid file on the disk.

I'm stuck on the pid, so either I'll get around to it at some point or someone else might sort it out. For now should be alright if I just commit the config file change?

ok, created the pull request. This is my first time doing any of this so I hope it's all good

Thanks for doing that.

no worries, benefits me as wel

how much disk space do I need to build omnios?