Re: how to increase space for tmpfs /tmp
On Wed, 28 Mar 2012 19:26:54 +0300, Andrei POPESCU wrote:
> On Mi, 28 mar 12, 15:12:19, Camaleón wrote:
>> On Wed, 28 Mar 2012 15:50:50 +0300, Andrei POPESCU wrote:
>>
>> > On Ma, 27 mar 12, 14:58:52, Camaleón wrote:
>> >>
>> >> IMO, the rule of thumb for applying a new default is asking
>> >> ourselves if the new default will cause any problem to the users. If
>> >> yes, then don't touch the old default and keep it the way it was.
>> >
>> > I don't agree.
>>
>> For any specific reason or just "don't agree" for the sake of not
>> agreeing? I'll be glad to hear your arguments on this :-)
>
> I thought I did below ;)
In this thread?
> but the short version would be "You can't make an omelette without
> breaking eggs"
Which explains little about your arguments (that's a general stanza) >:-)
>> >> If we are not going to get any improvement but just for the 10% of
>> >> our user-base, then we are failing the 90% of the rest.
>> >
>> > The improvement long term *could* be valuable enough to justify the
>> > pain. The correct way is usually not the easy way.
>>
>> And what (or who) decides what is "correct"?
>
> I think there is no correct answer that would apply anywhere[1]. It just
> happen that I agree with (most) of the choices taken by Debian and I say
> this by having followed all major discussions on -devel, -project and a
> few other mailing lists for several years now.
Debian is a big project and as such it has decided in many fields for the
defaults but again, that's not what we are talking here, this is about a
specific default which is now hitting some users.
> [1] except 42, of course :p
Sorry, I'm afraid I don't get that :-)
>> I've just seen another thread at this mailing list where "another" user
>> has been hit by this "correct" default. I don't mean that having "/tmp"
>> mounted as "tmpfs" is not correct but the default is clearly not suited
>> to many of the users as you can see.
>
> As far as I can tell /tmp as tmpfs is in testing since a few months and
> in unstable even longer. In my very humble opinion two threads in this
> period is hardly enough to call it "not suited to many of the users".
AFAIK, the discussion was started time ago at some of the "-devel"
mailing lists... and then it reached this general mailing list as soon as
wheezy users started having problems (me included). Although I'm using
wheezy I'm not usually following each "-devel" list, that will be too
much time for me, so I "discovered" the "out of space" problem when I was
uncompressing a "tar.gz" file. As I never enountered this message before,
I started a thread asking for feedback and found other users in the same
boat.
I can't count the exact number of users being hit by this but I'm sure
I'm not alone. To get quasi-accurate numbers a user targeted questionnary
could help to discover the maths behind this :-)
>> > One of the big reasons I love Debian is because it is not afraid to
>> > choose the hard path[1] when the long term benefits are worth it.
>>
>> (...)
>>
>> Although that's your personal opinion as you can easily understand it
>> has nothing to do with the issue we are currently debating. Every user
>> can/do love Debian for their own/different reasons but none of our
>> personal reasons can be used as arguments to make changes in the defaul
>> settings,
>
> I wasn't trying to use personal reasons as arguments, I was just trying
> to suggest that popularity or ease of implementation (or whatever) of a
> decision does not necessarily imply correctness and Debian has had the
> courage to make unpopular decisions or choose the harder to implement
> solution when it thought it was necessary.
Since when getting an "out of space" error message printed at your screen
and a faulty operation is something that can be considered "by popular
decision"? Of course not, so the own nature of the error is not something
I'd put inside a "feature" but more close to a "bug".
> I just happen to respect this in general, regardless if the decision
> proves to be "good" or "wrong".
Well, good decisions can enpower Debian (or any other piece of software
because this is not Debian specific, I'm afraid more distributions will
have to decide for this also...) while bad decisions can decrease its
users-base (which means they can consider another distributions).
>> unless there's a strong technical reason that proves they're the right
>> thing to do, which I still don't see for this specific issue.
>
> Well, isn't this your personal opinion (that there is no strong
> technical reason)? :)
My strong technical reason is all but personal: I failed to uncompress a
simple tar.gz file by this "correct" default, I never had experienced
this problem before. Nothing personal at all ;-)
Greetings,
--
Camaleón
Reply to: