nathans wrote:
> ----- Original Message -----
>> crash-resilience for systemd pmmgr/pmwebd
>
> This change needs to be reflected in the other daemons unit files too
> as was discussed awhile back, and the QA problems the initial attempt
> to do so introduced need to be resolved. We don't tend to do partial
> "piecemeal" fixes like this, because the rest tend to be forgotten or
> ignored, as has happened here.
> I reckon this needs to be finished, or reverted until done properly.
This is a contributor-hostile attitude. The fixes are obvious,
proper, self-contained, tested, working. You are proposing to hold
them hostage until some completely independent code is changed. If
you're thinking of a replay of my pmmgr container-subtarget code being
ransomed six months later by my fixes of "pmie -h", no, let's not do
that again.
There is no statement in HACKING that any random incremental
improvement in locality of the code base must - MUST! - be analogously
carried to every other nook & cranny. If it were a standard, it'd
have been violated at some point by every maintainer. If applied to
the community as a whole, it would dissuade changes. If it were a
standard applied selectively to only a few community members, no,
let's not do that either.
Incremental change is a good thing. Embrace it.
>> unresponsive-pmda pmie message: identify host
>> src/pmieconf/primary/pmda_status | 2
>
> This makes little sense (its a localhost-only rule) [...]
At the time the patch was written, it was not localhost-only. You did
not respond to later review comments at [1], so we don't know what
your intent is with respect to remote monitoring of pmda status. IMO
that remains a valuable service.
[1] http://oss.sgi.com/pipermail/pcp/2016-May/010697.html
- FChE
|