| 1 |
Duncan posted on Thu, 13 Mar 2014 13:45:14 +0000 as excerpted: |
| 2 |
|
| 3 |
> 2b) I wasn't satisfied with the default gpm unit file config either, but |
| 4 |
> by the time I got to that, I knew .includes were deprecated, and was in |
| 5 |
> the process of trying to figure out what the allusion to *.d/*.conf was |
| 6 |
> all about. |
| 7 |
> |
| 8 |
> But once I figured that out for random-seed, I was quickly able to add |
| 9 |
> /etc/systemd/system/gpm.service.d/execstart.conf , with a customized |
| 10 |
> ExecStart= line. =:^) |
| 11 |
|
| 12 |
Turned out that didn't work after all. I thought the modified ExecStart |
| 13 |
line would replace the previous one and thought it was working but I |
| 14 |
guess I hadn't tried a console mouse yet, as that didn't work. I was |
| 15 |
getting an error due to the two execstart lines where only one was |
| 16 |
allowed. |
| 17 |
|
| 18 |
So I ended up dup-ing the entire gpm.service file from /usr to /etc and |
| 19 |
editing the execstart line as necessary. |
| 20 |
|
| 21 |
> 5) A couple remaining minor service failures: |
| 22 |
> |
| 23 |
> 5a) ntp.service is still failing, for some reason. |
| 24 |
|
| 25 |
> Based on an aside-comment I read when researching something else, I |
| 26 |
> already have one lead to check on this one -- the comment said systemd's |
| 27 |
> privatetmp option doesn't play well with /var/tmp being a symlink to / |
| 28 |
> tmp, and that's what I have here (with /tmp in turn being tmpfs), and I |
| 29 |
> think the ntp unit might use privatetmp, so... But I've yet to confirm, |
| 30 |
> and then I'll have to look for alternatives, so... |
| 31 |
|
| 32 |
That was it. PrivateTmp doesn't like a symlinked /var/tmp. |
| 33 |
|
| 34 |
So I reset /var/tmp to a directory and setup an fstab entry to bind-mount |
| 35 |
the /tmp tmpfs to /var/tmp as well. Seems to be working at first blush, |
| 36 |
tho I haven't rebooted enough times yet to be sure there's no possible |
| 37 |
service dependency issues with it, but so far, it's working. =:^) |
| 38 |
|
| 39 |
|
| 40 |
> 5b) On openrc, I have a custom user-bootclean service that cleans up a |
| 41 |
> few stray user files at boot, that systemd created an attempted port |
| 42 |
> for, except it's failing. I've not investigated that one at all yet, |
| 43 |
> but just as I write this it occurs to me that the problem might be yet |
| 44 |
> another symlinked path mount-dependency issue, so I now have one lead to |
| 45 |
> investigate on that one, too. |
| 46 |
|
| 47 |
I haven't fully figured out why it's failing yet, but I think I can |
| 48 |
manage that with a tmpfiles.d entry. And openrc supports tmpfiles.d as |
| 49 |
well, so if I setup such an entry, openrc will hopefully handle that too, |
| 50 |
and I can simply delete the /etc/local.d/*.start file that's failing. |
| 51 |
That should be the clean modern solution in any case. |
| 52 |
|
| 53 |
But I'm too sleepy now to think straight, so I'll have to sleep on |
| 54 |
actually setting up and testing that one, lest I make a stupid mistake |
| 55 |
and end up deleting the entire user's homedir! |
| 56 |
|
| 57 |
-- |
| 58 |
Duncan - List replies preferred. No HTML msgs. |
| 59 |
"Every nonfree program has a lord, a master -- |
| 60 |
and if you use the program, he is your master." Richard Stallman |
Archives