| 1 |
On 31/08/2015 16:03, covici@××××××××××.com wrote: |
| 2 |
> Alan McKinnon <alan.mckinnon@×××××.com> wrote: |
| 3 |
> |
| 4 |
>> On 31/08/2015 13:49, covici@××××××××××.com wrote: |
| 5 |
>>>> A clue is in the ebuilds for systemd: |
| 6 |
>>>>> |
| 7 |
>>>>> sysv-utils? ( |
| 8 |
>>>>> !sys-apps/systemd-sysv-utils |
| 9 |
>>>>> !sys-apps/sysvinit ) |
| 10 |
>>>>> |
| 11 |
>>>>> That's a hard blocker, no way round it. It's in all the systemd ebuilds |
| 12 |
>>>>> for the current unstable versions. |
| 13 |
>>>>> |
| 14 |
>>>>> Do you have USE="sysv-utils" set for sysvinit? |
| 15 |
>>>>> |
| 16 |
>>>>> If so, to have both sysvinit and systemd, you will have to disable that |
| 17 |
>>>>> USE flag and see what comes next. |
| 18 |
>>> I put that use flag in there because I thought it would allow systemd to |
| 19 |
>>> generate a service from a script in /etc/init.d, but I will see what |
| 20 |
>>> happens when I remove that flag or maybe if there is another way to |
| 21 |
>>> accomplish that? |
| 22 |
>>> Well, that did it! It still is downgrading systemd, but that's not too |
| 23 |
>>> bad, thanks guys. |
| 24 |
>> |
| 25 |
>> $ euses -sf sysv-utils |
| 26 |
>> sys-apps/systemd:sysv-utils - Install sysvinit compatibility symlinks |
| 27 |
>> and manpages for init, telinit, halt, poweroff, reboot, runlevel, and |
| 28 |
>> shutdown |
| 29 |
>> |
| 30 |
>> |
| 31 |
>> That description is quite vague, and could mean many things. I'm no |
| 32 |
>> expert on systemd, but I would imagine that it already has it's own |
| 33 |
>> scripts to deal with those listed functions. I wonder what the use of |
| 34 |
>> the flag is then? Perhaps an old compatibility layer than is not needed now? |
| 35 |
>> |
| 36 |
>> |
| 37 |
>> I can't see a reason why systemd is being downgraded; the previous |
| 38 |
>> output either lists just "sys-apps/systemd" or uses a ">=" operator. |
| 39 |
>> Nothing to say why 219_p112 is the highest usable version. |
| 40 |
>> |
| 41 |
>> Once the emerge finishes and portage has done what it wants, run these |
| 42 |
>> commands: |
| 43 |
>> |
| 44 |
>> emerge -pv systemd |
| 45 |
>> emerge -pv =systemd-225 |
| 46 |
>> |
| 47 |
>> (225 being latest in the tree). Then we can see better why portage is |
| 48 |
>> doing what it does |
| 49 |
>> |
| 50 |
>> |
| 51 |
>> |
| 52 |
> |
| 53 |
> I think it has something to do with fail2ban -- the version of systemd |
| 54 |
> in the tree after the 219 version is 224-r1 and 225 and now portage is |
| 55 |
> saying |
| 56 |
> WARNING: One or more updates/rebuilds have been skipped due to a |
| 57 |
> dependency conflict: |
| 58 |
> and one of those says |
| 59 |
> (sys-apps/systemd-225:0/2::gentoo, ebuild scheduled for merge) |
| 60 |
> conflicts with^M |
| 61 |
> sys-apps/systemd[python(-),python_targets_python2_7(-),python_single_target_python2_7(+),python_targets_python3_4(-)] |
| 62 |
> required by (net-analyzer/fail2ban-0.9.3:0/0::gentoo, installed) |
| 63 |
> Does that make sense? |
| 64 |
> |
| 65 |
|
| 66 |
The words make sense, the meaning doesn't :-) |
| 67 |
|
| 68 |
It looks like fail2ban wants systemd without python support, but the |
| 69 |
true reason is still hidden. The fail2ban ebuild has this: |
| 70 |
|
| 71 |
RDEPEND=" |
| 72 |
... |
| 73 |
systemd? ( $(python_gen_cond_dep '|| ( |
| 74 |
dev-python/python-systemd[${PYTHON_USEDEP}] |
| 75 |
sys-apps/systemd[python(-),${PYTHON_USEDEP}] |
| 76 |
|
| 77 |
|
| 78 |
I'm thinking maybe you have a specific portage entry that's getting in |
| 79 |
the way. What are your results for: |
| 80 |
|
| 81 |
emerge --info |
| 82 |
grep -r python /etc/portage |
| 83 |
grep -r systemd /etc/portage |
| 84 |
|
| 85 |
|
| 86 |
-- |
| 87 |
Alan McKinnon |
| 88 |
alan.mckinnon@×××××.com |
Archives