Re: [gentoo-user] Partitioning strategy...?

[Alan McKinnon]

On Fri, 25 Nov 2011 20:53:17 +0700
Pandu Poluan  wrote:

> I want to build a Gentoo server box whose structure is
> highly-partitioned, like this:

partition setups are like lovers - highly variable. And the one that
suits you will suit almost no-one else.

Many of the recommendations you find on-line come from an earlier time
and the reason they got going is no longer valid for the most part. So
do take care to evaluate the real reason why you are doing something.

Valid reasons included:

You want to unmount a dir structure (/boot).
The fs type for a partition is different from that fs it mounts to
(often /var/log but these days most often used with tmpfs).
You need to mount an fs with different mount options to the fs it
mounts onto (/home noexec on multi-user setups for example)

The way to do this is not to search Google for recommendations, as
there is no such valid thing, but to figure out for yourself why you
want a mountpoint, calculate how much space *you* need, then do it.
Read other's experiences who use similar software as you by all means,
but that will be mere hints.

My own thoughts:

- I can't find a good reason anymore to have a local /usr separate. It's
  always mounted on my systems, even in maintenance mode (there's
  always at least one decent tool that the distro decided to put
  in /usr/sbin)

- /tmp is only useful on it's own if it's a tmpfs. Mine hasn't ever
  filled up anywhere (despite best efforts of users). tmpfs is general
  is an awesome idea.

- Keeping data and code separate is always a good idea. But only a few
  things in /var are critical like /var/log and /var/.
  Everything else is usually tiny and can safely live on /

- /boot is traditionally separate partly because long long long ago
  BIOSs couldn't read past 1024 cylinders which borked lilo. This is no
  longer true. 


> 
> /
> /boot
> /usr
> /tmp
> /usr/portage ==> via NFS
> /var
> /var/lib/postgresql
> /var/tmp
> /var/log
> /var/spool
> 
> (Not all of them will reside on the same physical disk; I have
> /dev/sda up to /dev/sdd)
> 
> I've been searching high and low for recommended numbers... and there
> are as many number as search-hits.
> 
> So. Care to share your partitioning strategy?
> 
> (And while we're at it, am I overdoing the partitioning?)
> 
> Rgds,



-- 
Alan McKinnnon
alan.mckin...@gmail.com

Re: [gentoo-user] 200MB waste from /usr/share/locale ?

[bill.longman]

Have you looked at localepurge?-- Sent from my HP TouchPadOn Nov 25, 2011 7:00 PM, Philip Webb  wrote: 26 Florian Philipp wrote:
> Am 26.11.2011 01:28, schrieb Sebastian Pipping:
>> It seems that /usr/share/locale is keeping files for many languages
>> not of any use to me: around 200MB in total.
>> Is there a way to configure this away that I am not aware of?
> Please follow the section "glibc Locales" in the Gentoo handbook.
> http://www.gentoo.org/doc/en/handbook/handbook-amd64.xml?full=1#book_part1_chap6

Yes, but that may not be the problem.  I have just  2  locales requested,
but  /etc/share/locale/  contains  90 MB  of material.
A quick look suggests that most of it is in subdirs LC_MESSAGES ,
wh seem to come from emerges at various dates;
they are binary, so I don't know what they were reporting.
Even under /usr/share/locale/en_US , which should show my  2  locales,
there are only such messages.  The locales do seem to work.

Can anyone explain what is going on ?

-- 
,,
SUPPORT ___//___,   Philip Webb
ELECTRIC   /] [] [] [] [] []|   Cities Centre, University of Toronto
TRANSIT`-O--O---'   purslowatchassdotutorontodotca

Re: [gentoo-user] 200MB waste from /usr/share/locale ?

[Mike Edenfield]

On 11/25/2011 10:00 PM, Philip Webb wrote:

26 Florian Philipp wrote:

Am 26.11.2011 01:28, schrieb Sebastian Pipping:

It seems that /usr/share/locale is keeping files for many languages
not of any use to me: around 200MB in total.
Is there a way to configure this away that I am not aware of?

Please follow the section "glibc Locales" in the Gentoo handbook.
http://www.gentoo.org/doc/en/handbook/handbook-amd64.xml?full=1#book_part1_chap6


Yes, but that may not be the problem.  I have just  2  locales requested,
but  /etc/share/locale/  contains  90 MB  of material.
A quick look suggests that most of it is in subdirs LC_MESSAGES ,
wh seem to come from emerges at various dates;
they are binary, so I don't know what they were reporting.
Even under /usr/share/locale/en_US , which should show my  2  locales,
there are only such messages.  The locales do seem to work.

Can anyone explain what is going on ?


Different packages include different levels of support for 
filtering their installed localization messages, typically 
one of "install everything", "install what's requested", or 
"whats a locale?"


The reason you mostly have files under LC_MESSAGES is 
because that's 99% of what is needed to localize a package. 
The files in there are string resource packages, 
translations of the strings used by the program, which are 
picked up by the localization library (gettext) 
automatically based on your locale settings. (coreutils 
installs file into LC_TIME for locales with date/time 
formatting requirements; I don't think I've ever seen any 
other locale files.)


The standard way to inform a package which languages you 
want is to set your LINGUAS variable in /etc/make.conf to 
the locale name(s) you want installed (without the charset 
specifier). LINGUAS works like any other portage expansion 
variables: for those packages that support it, you get a set 
of USE-flag-like language keywords set on build. (LINGUAS is 
the well-known environment variable used by most 
autotools-based packages to select languages, but portage 
provides support above and beyond that.)


Unfortunately, proper locale support is spotty -- mostly due 
to upstream maintainers being too lazy to properly add it to 
their builds. Instead, the package will install every 
message file it has available all the time.


You can safely delete any folders from /usr/share/locale for 
locales that you don't have installed, since the normal 
locale support in glibc will never ask for them. But they'll 
just get put back next time you upgrade the package.


--Mike

Re: [gentoo-user] Install problem - SATA CD-ROM drive (SOLVED)

[Dale]

Paul Colquhoun wrote:

On Fri, 25 Nov 2011 07:28:50 PM Albert W. Hopkins wrote:

On Sat, 2011-11-26 at 11:01 +1100, Paul Colquhoun wrote:

Some research tells me that this is a problem with SATA CD-ROMs, and
needs a
kernel configured with the libata module option atapi_enabled=1

Does anybody know of a Live CD that will work with a SATA CD drive?

AFAIK most (all) modern computers that have CDROM drives are SATA, at
least the one i bought a couple of years ago does.. so I'm wondering if
your problem is not that you have a SATA CDROM but something else.

Have you tried boot media other than the Gentoo livecd?


Actually I had tried 4 boot media *other* *than* the Gentoo livecd, for the
simple reason that I thought they had stopped making them!

Stupid me.

Looking at the Gentoo download page, they have an AMD64 minimal install CD
image created just days ago.

The Gentoo livecd works!  I should have looked there first and saved myself a
lot of trouble (and 4 blank CDs!).





Could it be something in the BIOS maybe?  My rig that I built last year 
has a SATA CD/DVD burner and it boots Gentoo CDs, Knoppix, systemrescue 
and another that I can't recall.  Knotix (?) or something.  It should 
work.  You may want to try systemrescue if you haven't already.  I'd 
figure this out before something happens and you have to boot something 
to fix it.


Dale

:-)  :-)

--
I am only responsible for what I said ... Not for what you understood or how 
you interpreted my words!

Re: [gentoo-user] 200MB waste from /usr/share/locale ?

[Philip Webb]

26 Florian Philipp wrote:
> Am 26.11.2011 01:28, schrieb Sebastian Pipping:
>> It seems that /usr/share/locale is keeping files for many languages
>> not of any use to me: around 200MB in total.
>> Is there a way to configure this away that I am not aware of?
> Please follow the section "glibc Locales" in the Gentoo handbook.
> http://www.gentoo.org/doc/en/handbook/handbook-amd64.xml?full=1#book_part1_chap6

Yes, but that may not be the problem.  I have just  2  locales requested,
but  /etc/share/locale/  contains  90 MB  of material.
A quick look suggests that most of it is in subdirs LC_MESSAGES ,
wh seem to come from emerges at various dates;
they are binary, so I don't know what they were reporting.
Even under /usr/share/locale/en_US , which should show my  2  locales,
there are only such messages.  The locales do seem to work.

Can anyone explain what is going on ?

-- 
,,
SUPPORT ___//___,   Philip Webb
ELECTRIC   /] [] [] [] [] []|   Cities Centre, University of Toronto
TRANSIT`-O--O---'   purslowatchassdotutorontodotca

Re: [gentoo-user] Partitioning strategy...?

[Pandu Poluan]

On Nov 26, 2011 9:05 AM, "Volker Armin Hemmann" 
wrote:
>
>
>
> 2011/11/25 Pandu Poluan 
>>
>> I want to build a Gentoo server box whose structure is
>> highly-partitioned, like this:
>>
>> /
>> /boot
>
>
> okay
>
>>
>> /usr
>
>
> why? just makes your life harder.
>

Why will it make my life harder.

>>
>> /tmp
>
>
> okay
>
>>
>> /usr/portage ==> via NFS
>
>
> if it makes you happy...
>

The portage tree will be shared among Gentoo boxen.

>>
>> /var
>
>
> okay
>
>>
>> /var/lib/postgresql
>
>
> ?
>

I'm using PostgreSQL, and the database IMO should be safely kept in a
separate partition.

>> /var/tmp
>
>
> okay
>
>>
>> /var/log
>
>
> why?
>

Postfix's log files are huge.

>> /var/spool
>
>
> why?

Postfix.

>>
>>
>> (Not all of them will reside on the same physical disk; I have
>> /dev/sda up to /dev/sdd)
>>
>> I've been searching high and low for recommended numbers... and there
>> are as many number as search-hits.
>>
>> So. Care to share your partitioning strategy?
>
>
> if you really have to split up /var like this, do yourself some favours
and spread it over several disks. Also don't put /var and /usr on the same
disk.
>

Indeed.

>>
>>
>> (And while we're at it, am I overdoing the partitioning?)
>
>
> yes, a lot.
>

:-)

Rgds,

Re: [gentoo-user] wicd and net-tools

[Allan Gottlieb]

On Fri, Nov 25 2011, Mark Knecht wrote:

> On Fri, Nov 25, 2011 at 2:58 PM, Allan Gottlieb  wrote:
>> In the last day or two wicd broke badly due to a net-tools upgrade.
>>
>> The recommended workarounds are to specify USE=old-output
>> for net-tools or to downgrade net-tools one version (I am ~amd64).
>>
>> Neither of these have helped me.  Wicd cannot start either the
>> wired or wireless interface.
>>
>> When I type
>> killall dhcpd; dhcpd
>>
>> Both networks come up but the wireless gets many errors (as reported
>> by ifconfig).  Also sometimes the network goes down again.
>>
>> Two days ago both networks were solid.
>>
>> Another machine that I was in the process of bringing up (gentoo fully
>> installed from the handbook, but many utilities not yet done) also fails
>> under wicd today (worked fine two days ago).  For this machine
>> killall dhcp; dhcp
>> reports that "no interfaces have a carrier"
>> The machine is 10 feet from the router/wap (linksys) and again
>> all was well 2 days ago.
>>
>> Help would be very much appreciated.
>>
>> thanks,
>> allan
>>
>> PS I couldn't connect yesterday and today, the new gnome (3.2) hit
>> testing.  Since I don't want to bite that big, potentially disruptive
>> update with this network problem (the overlay version failed for me), I
>> have not done an update world for 2 or 3 days.
>>
>>
>
> Not to be too glib but isn't this sort of problem exactly what the
> whole stable vs ~amd64 decision is really all about?
>
> Anyway, my machines are stable with a few ~amd64 entries in
> package.keywords. I don't keyword either of these packages and I have
> wicd-1.7.1_beta2-r4 & net-tools-1.60_p20110409135728 on my laptop.
> Both wireless & wired come up fine here.
>
> I have dhcpd on the machine but I don't run it.
>
> Not exactly sure how any of that will help you but I'm fully up to
> date as of today and everything seems to be working fine.
>
> - Mark

Right.  I am running testing.  But so are many on this list and I
believe many of those use wicd.  I was hoping someone would have solved
this same problem.

allan

Re: [gentoo-user] 200MB waste from /usr/share/locale ?

[Florian Philipp]

Am 26.11.2011 01:28, schrieb Sebastian Pipping:
> Hello!
> 
> 
> It seems that /usr/share/locale is keeping files for many languages not
> of any use to me: around 200MB in total.
> 
> Is there a way to configure this away that I am not aware of?
> 
> Thanks in advance,
> 
> 
> 
> Sebastian
> 

Please follow the section "glibc Locales" in the Gentoo handbook.
http://www.gentoo.org/doc/en/handbook/handbook-amd64.xml?full=1#book_part1_chap6

Regards,
Florian Philipp



signature.asc
Description: OpenPGP digital signature

Re: [gentoo-user] Partitioning strategy...?

[Volker Armin Hemmann]

2011/11/25 Pandu Poluan 

> I want to build a Gentoo server box whose structure is
> highly-partitioned, like this:
>
> /
> /boot
>

okay


> /usr
>

why? just makes your life harder.


> /tmp
>

okay


> /usr/portage ==> via NFS
>

if it makes you happy...


> /var
>

okay


> /var/lib/postgresql
>

?

/var/tmp
>

okay


> /var/log
>

why?

/var/spool
>

why?

>
> (Not all of them will reside on the same physical disk; I have
> /dev/sda up to /dev/sdd)
>
> I've been searching high and low for recommended numbers... and there
> are as many number as search-hits.
>
> So. Care to share your partitioning strategy?
>

if you really have to split up /var like this, do yourself some favours and
spread it over several disks. Also don't put /var and /usr on the same
disk.


>
> (And while we're at it, am I overdoing the partitioning?)
>

yes, a lot.

Re: [gentoo-user] Partitioning strategy...?

[Pandu Poluan]

On Nov 26, 2011 5:05 AM, "Neil Bothwick"  wrote:
>
> On Fri, 25 Nov 2011 23:12:42 +0700, Pandu Poluan wrote:
>
> > > I don't use LVM but I suspect that on this list that would be the #1
> > > recommendation to take care of the numbers question.
>
> > Everytime I read some guide on LVM, my eyes becomes blurry, the room
> > starts spinning, and I can hear wolves howling ... :D
> >
> > Seriously, LVM looks mighty nice, but it also looks (and is!) mighty
> > complex.
>
> It may look it, but you only have to learn the concepts once. That many
> physical partitions will be extra work forever, imagine what happens when
> one of the middle ones is no longer big enough.
>
> Seriously, spend half an hour reading up on LVM and you'll never regret
> it.
>

Oh, I certainly will. But not now. I need to get this box up, staged, and
productioned ASAP.

Afterwards, I'll commit myself to understanding LVM.

Rgds,

Re: [gentoo-user] Install problem - SATA CD-ROM drive (SOLVED)

[Paul Colquhoun]

On Fri, 25 Nov 2011 07:28:50 PM Albert W. Hopkins wrote:
> On Sat, 2011-11-26 at 11:01 +1100, Paul Colquhoun wrote:
> > Some research tells me that this is a problem with SATA CD-ROMs, and
> > needs a
> > kernel configured with the libata module option atapi_enabled=1
> > 
> > Does anybody know of a Live CD that will work with a SATA CD drive?
> 
> AFAIK most (all) modern computers that have CDROM drives are SATA, at
> least the one i bought a couple of years ago does.. so I'm wondering if
> your problem is not that you have a SATA CDROM but something else.
> 
> Have you tried boot media other than the Gentoo livecd?


Actually I had tried 4 boot media *other* *than* the Gentoo livecd, for the 
simple reason that I thought they had stopped making them!

Stupid me.

Looking at the Gentoo download page, they have an AMD64 minimal install CD 
image created just days ago.

The Gentoo livecd works!  I should have looked there first and saved myself a 
lot of trouble (and 4 blank CDs!).


-- 
Reverend Paul Colquhoun, ULC.http://andor.dropbear.id.au/~paulcol
 Before you criticize someone, you should walk a mile in their shoes.
Then, when you do, you'll be a mile away, and you'll have their shoes.

Re: [gentoo-user] wicd and net-tools

[Mark Knecht]

On Fri, Nov 25, 2011 at 2:58 PM, Allan Gottlieb  wrote:
> In the last day or two wicd broke badly due to a net-tools upgrade.
>
> The recommended workarounds are to specify USE=old-output
> for net-tools or to downgrade net-tools one version (I am ~amd64).
>
> Neither of these have helped me.  Wicd cannot start either the
> wired or wireless interface.
>
> When I type
> killall dhcpd; dhcpd
>
> Both networks come up but the wireless gets many errors (as reported
> by ifconfig).  Also sometimes the network goes down again.
>
> Two days ago both networks were solid.
>
> Another machine that I was in the process of bringing up (gentoo fully
> installed from the handbook, but many utilities not yet done) also fails
> under wicd today (worked fine two days ago).  For this machine
> killall dhcp; dhcp
> reports that "no interfaces have a carrier"
> The machine is 10 feet from the router/wap (linksys) and again
> all was well 2 days ago.
>
> Help would be very much appreciated.
>
> thanks,
> allan
>
> PS I couldn't connect yesterday and today, the new gnome (3.2) hit
> testing.  Since I don't want to bite that big, potentially disruptive
> update with this network problem (the overlay version failed for me), I
> have not done an update world for 2 or 3 days.
>
>

Not to be too glib but isn't this sort of problem exactly what the
whole stable vs ~amd64 decision is really all about?

Anyway, my machines are stable with a few ~amd64 entries in
package.keywords. I don't keyword either of these packages and I have
wicd-1.7.1_beta2-r4 & net-tools-1.60_p20110409135728 on my laptop.
Both wireless & wired come up fine here.

I have dhcpd on the machine but I don't run it.

Not exactly sure how any of that will help you but I'm fully up to
date as of today and everything seems to be working fine.

- Mark

Re: [gentoo-user] 200MB waste from /usr/share/locale ?

[Albert W. Hopkins]

On Sat, 2011-11-26 at 01:28 +0100, Sebastian Pipping wrote:
> It seems that /usr/share/locale is keeping files for many languages
> not
> of any use to me: around 200MB in total.
> 
> Is there a way to configure this away that I am not aware of?
> 
> Thanks in advance, 

What do you have in LINGUAS and locales.conf?  My /usr/share/locale is
only about 16MB on a GNOME desktop, 9MB on a base install and about 26MB
on a KDE VM (not that 200MB is really that big by today's standards).

-a

Re: [gentoo-user] Install problem - SATA CD-ROM drive

[Albert W. Hopkins]

On Sat, 2011-11-26 at 11:01 +1100, Paul Colquhoun wrote:
> Some research tells me that this is a problem with SATA CD-ROMs, and
> needs a 
> kernel configured with the libata module option atapi_enabled=1
> 
> Does anybody know of a Live CD that will work with a SATA CD drive? 

AFAIK most (all) modern computers that have CDROM drives are SATA, at
least the one i bought a couple of years ago does.. so I'm wondering if
your problem is not that you have a SATA CDROM but something else.

Have you tried boot media other than the Gentoo livecd?

[gentoo-user] 200MB waste from /usr/share/locale ?

[Sebastian Pipping]

Hello!


It seems that /usr/share/locale is keeping files for many languages not
of any use to me: around 200MB in total.

Is there a way to configure this away that I am not aware of?

Thanks in advance,



Sebastian

[gentoo-user] Re: experience with rsnapshot

[Grant Edwards]

On 2011-11-25, cov...@ccs.covici.com  wrote:

> I noticed that there was no real restore,

Um, it's a regular file sysmte, so you use "cp -a" to restore.

> but as you say you can usually find what you are looking for.  I will
> probably try on an experimental basis.

-- 
Grant

Re: [gentoo-user] Partitioning strategy...?

[Dale]

Mark Knecht wrote:

On Fri, Nov 25, 2011 at 11:51 AM, Walter Dnes  wrote:

On Fri, Nov 25, 2011 at 10:17:07AM -0800, Mark Knecht wrote

On Fri, Nov 25, 2011 at 9:53 AM, Pandu Poluan  wrote:


/mnt/distfiles ext3/4 (shared via nfs)


Why do you separate the distfiles from the portage tree?

Rgds,

distfiles has a tendency to grow large over the years. IIRC nothing
cleans it up automatically so having it separate is just a simple
safety mechanism to not run out of disk space after emerge -fDuN
@world, etc.

  man eclean
http://gpio.ca/cgi-bin/man/man2html?1+eclean

--
Walter Dnes


Yes, true, but all I said was that TTBOMK nothing does it
automatically, not that it cannot be automated.

- Mark




I use http-replicator and I wish it would clean distfiles from it 
instead of /usr/portage/distfiles.  When I run repcacheman, it cleans 
out distfiles already.  I just need to clean up http-rep's directory.  
Right now, I do a emerge -ef world, rm http-rep's stuff then run 
repcacheman again.  Thing is, it cleans out all the stuff my x86 box 
needs that my amd64 box doesn't since I always forget to run emerge -ef 
world on the x86 box too. < sighs >


Dale

:-)  :-)

--
I am only responsible for what I said ... Not for what you understood or how 
you interpreted my words!

[gentoo-user] Install problem - SATA CD-ROM drive

[Paul Colquhoun]

I've just started building a new system from scratch, and have run into a 
problem.

The motherboard only has SATA and USB conectors, no legacy IDE/PATA/FDD/etc 
connectors at all.

The problem is that all 4 of the Live CDs I have tried fail to boot 
completely.

All of them bring up the initial menus, then run the kernel, but then fail at 
the same point, which is when they attempt to (re)mount the filesystems from 
the CD, which they all claim not to be able to find.

Selecting a shell at this point shows that there are no /dev/sr? devices 
detected, and all the /dev/sd? devices correspont to hard drives.

Some research tells me that this is a problem with SATA CD-ROMs, and needs a 
kernel configured with the libata module option atapi_enabled=1

Does anybody know of a Live CD that will work with a SATA CD drive?

The other option I am exploring is borrowing a USB CD drive to see if that 
will work. 

All I need it for is the initial bootstrap into Gentoo.

With the growing trend of SATA-only motherboards, I hope there is at least 1 
Live CD out there that works.

The Live CD versions that I have tried are:
Sabayon Awsome 7
Kubuntu 11.10
Fedora-16
SystemRescueCD 2.4.0

P.S.  This is an AMD64 motherboard, so a 64 bit kernel is needed.

Thanks.


-- 
Reverend Paul Colquhoun, ULC.http://andor.dropbear.id.au/~paulcol
 Before you criticize someone, you should walk a mile in their shoes.
Then, when you do, you'll be a mile away, and you'll have their shoes.

Re: [gentoo-user] Partitioning strategy...?

[Dale]

Neil Bothwick wrote:

On Fri, 25 Nov 2011 23:12:42 +0700, Pandu Poluan wrote:


I don't use LVM but I suspect that on this list that would be the #1
recommendation to take care of the numbers question.

Everytime I read some guide on LVM, my eyes becomes blurry, the room
starts spinning, and I can hear wolves howling ... :D

Seriously, LVM looks mighty nice, but it also looks (and is!) mighty
complex.

It may look it, but you only have to learn the concepts once. That many
physical partitions will be extra work forever, imagine what happens when
one of the middle ones is no longer big enough.

Seriously, spend half an hour reading up on LVM and you'll never regret
it.





Besides, even I can use LVM now.  I even reduced one and took a drive 
off, mostly to learn.  Between Neil and Alan, plus others, you would 
have some good helpers here.  Heck, read over my old threads.  I even 
posted the commands I used to remove a drive a few days ago.  There's 
not much better than having someone that has already done it to post the 
commands used.  At least then you know it works.


Am I going to put / on LVM, not yet.  I'm still a can or two short of a 
six pack.  I got to get the init thingy to work and work WELL first.


If I were you, I would at least try to put /boot and / outside LVM then 
everything else on LVM.  Just make sure /boot and / have PLENTY of space 
since they are pretty much committed at that point.  This is something I 
am thinking of doing on my rig and one reason I removed the a drive from 
LVM.  I needed some space to swap things around even from a CD/DVD boot.


Once you learn how to use it, it really is nice.  Setting it up is not 
bad at all.  It's when you need to move data that you can't back up or 
afford to lose that it gets hairy.  That is true for traditional 
partitions to tho.


Dale

:-)  :-)

--
I am only responsible for what I said ... Not for what you understood or how 
you interpreted my words!

Re: [gentoo-user] Partitioning strategy...?

[Dale]

Neil Bothwick wrote:

On Fri, 25 Nov 2011 14:07:48 -0600, Dale wrote:


Here is one thing to think about on LVM.  If you put /usr on a separate
partition, you will need the init thingy, thanks to the dev at fedora.

Not yet you don't. I'm happily running a separate /usr (LVM has nothing
to do with it) without an initramfs. On my new box, I do have an
initramfs, but that's because I have / on LVM, soon to be on an encrypted
volume.





That was what I meant to say.  If he has / on LVM, he needs a init 
thingy even now.


You don't need the init thingy for a separate /usr now but we all know 
it is coming.  I wouldn't want him to do a nice install then run into a 
known coming issue that throws things into a mess.  Heck, even I don't 
want to reinstall just for the heck of it.  I don't many who does.


Thanks for clarifying my point tho.  Maybe I do need to drink coffee 
when I first get up.  :/


Dale

:-)  :-)

--
I am only responsible for what I said ... Not for what you understood or how 
you interpreted my words!

Re: [gentoo-user] glibc-2.14.1 upgrade

[Volker Armin Hemmann]

2011/11/25 Florian Philipp 

> Am 25.11.2011 19:11, schrieb Pandu Poluan:
> >
> > On Nov 26, 2011 12:05 AM, "微菜"  > > wrote:
> >>
> >> 于 2011年11月24日 15:34, justin 写道:
> >> > On 24/11/11 06:27, Paul Hartman wrote:
> >> >> Hi,
> >> >>
> >> >> After emerging glibc-2.14.1 today, pam stopped working, which
> >> >> prevented KDE from working and some other things. I got this kind of
> >> >> message:
> >> >>
> >> >> /lib64/libc.so.6: version `GLIBC_2.14' not found (required by
> >> >> /lib64/libcrypt.so.1)
> >> >>
> >> >> There were no @preserved-rebuild and revdep-rebuild found nothing. I
> >> >> rebuilt pam and things seem to be working again. Are there any other
> >> >> packages I should rebuild before encountering a problem? Or some way
> >> >> to detect which need to be rebuilt? Should I re-emerge world against
> >> >> my new glibc? :)
> >> >>
> >> >> Thanks,
> >> >> Paul
> >> >>
> >> >
> >> > Hi Paul,
> >> >
> >> > after an glibc upgrade it is time to do an
> >> >
> >> > emerge -e system world
> >>
> >> you must be kidding me . it will take days to complete.
> >>
> >
> > On my virtualized servers, emerge -e @system @world, a total of
> > 170-something packages, takes only about 6-7 hours. Remember, these are
> > VMs, which means the 2 vCPUs I assigned them are actually shared with
> > other VMs.
> >
> > On a non-virtualized system, should be faster. Of course, if you have
> > sizable packages, it might take around 24 hours to complete.
> >
> > Rgds,
> >
>
> I've used Gentoo since 2006 and never had any reason to emerge -e. I can
> also cite posts on this mailing list going back as far as 2007 (just my
> local archive, mind you) that say it is unnecessary for glibc updates.
> Or, to cite Volker Armin Hemmann's eloquent reply from 08-11-2007: "no" ;-)
>
> Regards,
> Florian Philipp
>
>
sys-libs/glibc
 Available versions:  (2.2) (~)2.9_p20081201-r3!s 2.10.1-r1!s 2.11.3!s
(~)2.12.1-r3!s 2.12.2!s (~)2.13-r2!s (~)2.13-r4!s (~)2.14!s
(~)2.14.1!s{tbz2} **!s

{crosscompile_opts_headers-only debug gd glibc-omitfp hardened
multilib nls profile selinux vanilla}
 Installed versions:  2.14.1(2.2)!s{tbz2}(18:07:50 23.11.2011)(gd
glibc-omitfp multilib nls -crosscompile_opts_headers-only -debug -hardened
-profile -selinux -vanilla)

no -e world

login working fine. Posting from KDE 4.7.80

so, I stay with the 'no' from back than.

you don't need to do an -e world on glibc or header updates.

You CAN do it, to make use of some new and awesome feature you might hope
to be there. But for almost all people it is just a waste of time and
energy.

[gentoo-user] wicd and net-tools

[Allan Gottlieb]

In the last day or two wicd broke badly due to a net-tools upgrade.

The recommended workarounds are to specify USE=old-output
for net-tools or to downgrade net-tools one version (I am ~amd64).

Neither of these have helped me.  Wicd cannot start either the
wired or wireless interface.

When I type
killall dhcpd; dhcpd

Both networks come up but the wireless gets many errors (as reported
by ifconfig).  Also sometimes the network goes down again.

Two days ago both networks were solid.

Another machine that I was in the process of bringing up (gentoo fully
installed from the handbook, but many utilities not yet done) also fails
under wicd today (worked fine two days ago).  For this machine
killall dhcp; dhcp
reports that "no interfaces have a carrier"
The machine is 10 feet from the router/wap (linksys) and again
all was well 2 days ago.

Help would be very much appreciated.

thanks,
allan

PS I couldn't connect yesterday and today, the new gnome (3.2) hit
testing.  Since I don't want to bite that big, potentially disruptive
update with this network problem (the overlay version failed for me), I
have not done an update world for 2 or 3 days.

Re: [gentoo-user] glibc-2.14.1 upgrade

[Michael Mol]

106 minutes, 177 packsges on a fresh install (on a dual E5345 I just got my
hands on). God help you if you if you have KDE installed, though. Even
without it, my core desktop has somewhere between 500-700 packages. Builds
overnight on my Phenom 9650.

ZZ
On Nov 25, 2011 2:04 PM, "Florian Philipp"  wrote:

> Am 25.11.2011 19:11, schrieb Pandu Poluan:
> >
> > On Nov 26, 2011 12:05 AM, "微菜"  > > wrote:
> >>
> >> 于 2011年11月24日 15:34, justin 写道:
> >> > On 24/11/11 06:27, Paul Hartman wrote:
> >> >> Hi,
> >> >>
> >> >> After emerging glibc-2.14.1 today, pam stopped working, which
> >> >> prevented KDE from working and some other things. I got this kind of
> >> >> message:
> >> >>
> >> >> /lib64/libc.so.6: version `GLIBC_2.14' not found (required by
> >> >> /lib64/libcrypt.so.1)
> >> >>
> >> >> There were no @preserved-rebuild and revdep-rebuild found nothing. I
> >> >> rebuilt pam and things seem to be working again. Are there any other
> >> >> packages I should rebuild before encountering a problem? Or some way
> >> >> to detect which need to be rebuilt? Should I re-emerge world against
> >> >> my new glibc? :)
> >> >>
> >> >> Thanks,
> >> >> Paul
> >> >>
> >> >
> >> > Hi Paul,
> >> >
> >> > after an glibc upgrade it is time to do an
> >> >
> >> > emerge -e system world
> >>
> >> you must be kidding me . it will take days to complete.
> >>
> >
> > On my virtualized servers, emerge -e @system @world, a total of
> > 170-something packages, takes only about 6-7 hours. Remember, these are
> > VMs, which means the 2 vCPUs I assigned them are actually shared with
> > other VMs.
> >
> > On a non-virtualized system, should be faster. Of course, if you have
> > sizable packages, it might take around 24 hours to complete.
> >
> > Rgds,
> >
>
> I've used Gentoo since 2006 and never had any reason to emerge -e. I can
> also cite posts on this mailing list going back as far as 2007 (just my
> local archive, mind you) that say it is unnecessary for glibc updates.
> Or, to cite Volker Armin Hemmann's eloquent reply from 08-11-2007: "no" ;-)
>
> Regards,
> Florian Philipp
>
>

Re: [gentoo-user] glibc-2.14.1 upgrade

[Albert W. Hopkins]

On Fri, 2011-11-25 at 22:30 +, Neil Bothwick wrote:
> On Fri, 25 Nov 2011 19:59:51 +0100, Florian Philipp wrote:
> 
> > I've used Gentoo since 2006 and never had any reason to emerge -e.
> 
> Apart from the occasional gcc ABI change, the only time I do it is after
> a fresh install, so that everything is built with the installed toolchain.

I do it about once per month, just to make sure everthing works as
should, and, if not, usually I'll submit a bug report if necessary.

Re: [gentoo-user] glibc-2.14.1 upgrade

[Neil Bothwick]

On Fri, 25 Nov 2011 19:59:51 +0100, Florian Philipp wrote:

> I've used Gentoo since 2006 and never had any reason to emerge -e.

Apart from the occasional gcc ABI change, the only time I do it is after
a fresh install, so that everything is built with the installed toolchain.


-- 
Neil Bothwick

In possession of a mind not merely twisted, but actually sprained.


signature.asc
Description: PGP signature

Re: [gentoo-user] Error message for several ebuilds when updating world.

[Neil Bothwick]

On Fri, 25 Nov 2011 16:58:56 -0500, Walter Dnes wrote:

> * the rest of the drive is one large reiserfs partition
> 
>   The power of LVM, without the pain.

Hardly. For instance, a runaway program filling logfiles will
fill up everything, you can't isolate, for example, personal and system
data.

I don't see how you have use your method with multiple drives without
serious kludging.

LVM is about a lot more than keeping everything on one partition.


-- 
Neil Bothwick

Monday is the root of all evil!


signature.asc
Description: PGP signature

Re: [gentoo-user] Error message for several ebuilds when updating world.

[Walter Dnes]

On Fri, Nov 25, 2011 at 02:27:04PM -0600, Dale wrote

> The funny thing, I have to leave mine on tmpfs because when I update 
> LOo, I don't have enough space on /var.  I have enough ram tho.  lol

  I know I have enough diskspace.  See my reply to Pandu in the
"Partitioning strategy...?" thread.

* 250 megabyte / partition (ext2fs)
* 4 gigs of swap
* the rest of the drive is one large reiserfs partition

  The power of LVM, without the pain.

-- 
Walter Dnes 

Re: [gentoo-user] Partitioning strategy...?

[Neil Bothwick]

On Sat, 26 Nov 2011 00:53:57 +0700, Pandu Poluan wrote:

> Why do you separate the distfiles from the portage tree?

A better question is why put the distfiles in the middle of the portage
tree? It really makes no sense, they are two very different types of data.

A separate DISTDIR also makes sense when sharing it over NFS between
various  machines.


-- 
Neil Bothwick

If at first you don't succeed, call in an airstrike.


signature.asc
Description: PGP signature

Re: [gentoo-user] Partitioning strategy...?

[Neil Bothwick]

On Fri, 25 Nov 2011 14:07:48 -0600, Dale wrote:

> Here is one thing to think about on LVM.  If you put /usr on a separate 
> partition, you will need the init thingy, thanks to the dev at fedora. 

Not yet you don't. I'm happily running a separate /usr (LVM has nothing
to do with it) without an initramfs. On my new box, I do have an
initramfs, but that's because I have / on LVM, soon to be on an encrypted
volume.


-- 
Neil Bothwick

Assassins do it from behind.


signature.asc
Description: PGP signature

Re: [gentoo-user] Partitioning strategy...?

[Neil Bothwick]

On Fri, 25 Nov 2011 23:12:42 +0700, Pandu Poluan wrote:

> > I don't use LVM but I suspect that on this list that would be the #1
> > recommendation to take care of the numbers question.

> Everytime I read some guide on LVM, my eyes becomes blurry, the room
> starts spinning, and I can hear wolves howling ... :D
> 
> Seriously, LVM looks mighty nice, but it also looks (and is!) mighty
> complex.

It may look it, but you only have to learn the concepts once. That many
physical partitions will be extra work forever, imagine what happens when
one of the middle ones is no longer big enough.

Seriously, spend half an hour reading up on LVM and you'll never regret
it.


-- 
Neil Bothwick

RAM disk is *not* an installation procedure.


signature.asc
Description: PGP signature

Re: [gentoo-user] Re: Partitioning strategy...?

[Neil Bothwick]

On Fri, 25 Nov 2011 20:55:02 +0700, Pandu Poluan wrote:

> I also wonder if /var/tmp can be shared between boxen (via NFS),
> assuming I ensure that no two boxen perform emerge at the same time...

That will really slow down emerges.


-- 
Neil Bothwick

Psychiatrists say that 1 of 4 people are mentally ill.
Check three friends. If they're OK, you're it.


signature.asc
Description: PGP signature

[gentoo-user] Anti-aliasing, hinting or whatever.

[Stayvoid]

Hi there!

I have some problems with several fonts.
For example, "0" in Andale Mono is displayed with a dot shifted by
several points.
This shift looks different in different sizes.
How to fix this?

Cheers.

[gentoo-user] Fonts installation.

[Stayvoid]

Hi there!

I have some fonts which are not included in the repository.
How can I install them?

Cheers!

Re: [gentoo-user] Partitioning strategy...?

[Stéphane Guedon]

On Friday 25 November 2011 21:09:04 Mark Knecht wrote:
> On Fri, Nov 25, 2011 at 11:51 AM, Walter Dnes  wrote:
> > On Fri, Nov 25, 2011 at 10:17:07AM -0800, Mark Knecht wrote
> > 
> >> On Fri, Nov 25, 2011 at 9:53 AM, Pandu Poluan  wrote:
> >> 
> >> 
> >> >> /mnt/distfiles ext3/4 (shared via nfs)
> >> > 
> >> > Why do you separate the distfiles from the portage tree?
> >> > 
> >> > Rgds,
> >> 
> >> distfiles has a tendency to grow large over the years. IIRC nothing
> >> cleans it up automatically so having it separate is just a simple
> >> safety mechanism to not run out of disk space after emerge -fDuN
> >> @world, etc.
> > 
> >  man eclean
> > http://gpio.ca/cgi-bin/man/man2html?1+eclean
> > 
> > --
> > Walter Dnes 
> 
> Yes, true, but all I said was that TTBOMK nothing does it
> automatically, not that it cannot be automated.
> 
> - Mark

I find something on forums.gentoo.org that is called distfiles-cleanup. it's a 
perl scrit to clean distfiles by release order. Search it.

-- 
Stéphane Guedon
page web : http://www.22decembre.eu/
carte de visite : http://www.22decembre.eu/downloads/Stephane-Guedon.vcf
clé publique gpg : http://www.22decembre.eu/downloads/Stephane-Guedon.asc


signature.asc
Description: This is a digitally signed message part.

Re: [gentoo-user] Error message for several ebuilds when updating world.

[Dale]

Walter Dnes wrote:

On Fri, Nov 25, 2011 at 01:31:45AM -0600, Dale wrote


http://forums.gentoo.org/viewtopic-p-6878530.html?sid=924b3c65e5584c4dc81672583d97b85d#6878530

You need more ram, swap space or maybe drive space for the portage work
directory.  Basically, it's running out of space somewhere.

   Thanks for the pointer.  I think I've figured it it out now.  When I
set up my laptop, I copied most of /etc/make.conf verbatim from my
desktop, including the line...
PORTAGE_TMPFS="/dev/shm"

   That works just fine on my desktop with 8 gigs of ram.  Unfortunately,
my laptop only has 4 gigs, so that doesn't work quite so well.  I've
changed that line to...
PORTAGE_TMPFS="/tmp"

   Builds will be slower, no doubt.



I put the portage work directory on tmpfs in ram and I can't tell much 
difference.  When I tested it, having portage's work directory on tmpfs 
was actually a bit slower, just by a few seconds tho.  So, having it on 
a drive doesn't seem to matter much.  I guess drives are pretty fast 
nowadays.


The funny thing, I have to leave mine on tmpfs because when I update 
LOo, I don't have enough space on /var.  I have enough ram tho.  lol


Dale

:-)  :-)

--
I am only responsible for what I said ... Not for what you understood or how 
you interpreted my words!

Re: [gentoo-user] Re: Partitioning strategy...?

[Dale]

Francesco Talamona wrote:
IMO /usr/portage and /boot are too small, these are the respective 
sizes in my system: [root@aemaeth:~]$ du -sh /usr/portage 19G 
/usr/portage [root@aemaeth:~]$ du -sh /boot 52M /boot HTH Francesco 


This is my size and it works fine with the occasional use of eclean:

/dev/sda6 11535344   5389288   5560088  50% /usr/portage

It's a 12Gb partition and only half full.  That is with a full install 
of KDE4, LOo and Fluxbox for back-up.


Dale

:-)  :-)

--
I am only responsible for what I said ... Not for what you understood or how 
you interpreted my words!

Re: [gentoo-user] Partitioning strategy...?

[Dale]

Dale wrote:


Oh, here is a funny one.  Imagine walking up to the computer and 
seeing knotify taking up 14Gbs of ram.  O_O  My rig was using all the 
ram, some cache and slow as leap year.  I kill -9'd that thing.  G!!


Dale

:-)  :-)


Instead of *cache*, make that *swap*.  What was I thinking?  Oh, still 
sort of pissed at fedora over the init thingy.  That's it.


Dale

:-)  :-)

--
I am only responsible for what I said ... Not for what you understood or how 
you interpreted my words!

Re: [gentoo-user] Partitioning strategy...?

[Mark Knecht]

On Fri, Nov 25, 2011 at 11:51 AM, Walter Dnes  wrote:
> On Fri, Nov 25, 2011 at 10:17:07AM -0800, Mark Knecht wrote
>> On Fri, Nov 25, 2011 at 9:53 AM, Pandu Poluan  wrote:
>> 
>> >> /mnt/distfiles ext3/4 (shared via nfs)
>> >>
>> >
>> > Why do you separate the distfiles from the portage tree?
>> >
>> > Rgds,
>>
>> distfiles has a tendency to grow large over the years. IIRC nothing
>> cleans it up automatically so having it separate is just a simple
>> safety mechanism to not run out of disk space after emerge -fDuN
>> @world, etc.
>
>  man eclean
> http://gpio.ca/cgi-bin/man/man2html?1+eclean
>
> --
> Walter Dnes 
>

Yes, true, but all I said was that TTBOMK nothing does it
automatically, not that it cannot be automated.

- Mark

Re: [gentoo-user] Partitioning strategy...?

[Walter Dnes]

On Fri, Nov 25, 2011 at 11:12:42PM +0700, Pandu Poluan wrote

> Everytime I read some guide on LVM, my eyes becomes blurry, the room
> starts spinning, and I can hear wolves howling ... :D
> 
> Seriously, LVM looks mighty nice, but it also looks (and is!) mighty complex.
> 
> So, I want to start from something simple.
> 
> Aaaanyways, after reviewing my production boxes, I decided to
> implement the following strategy:
> 
> /  == 800 MiB
> /boot == 20 MiB
> /usr == 1800 MiB
> /usr/portage == 2000 MiB
> /var == 4000 MiB
> /var/lib/postgresql == 1000 MiB
> 
> Comments, suggestions, are welcome :)

  I have my own weird approach that's even weirder than my mdev setup.
I start with...
* 250 megabytes for / as ext2fs (No that is not a typo)
* 4 gigs for swap
* the rest of the drive is /home as one huge reiserfs partition

  And I do *NOT* use LVM.  "fdisk -l" shows...

   Device Boot  Start End  Blocks   Id  System
/dev/sda12048   976773167   4883855605  Extended
/dev/sda54096  516095  256000   83  Linux
/dev/sda6  518144 8906751 4194304   83  Linux
/dev/sda7 8908800   976773167   483932184   83  Linux

"df" shows

Filesystem   1K-blocks  Used Available Use% Mounted on
rootfs  247919 29315205804  13% /
/dev/root   247919 29315205804  13% /
devtmpfs 10240 0 10240   0% /dev
rc-svcdir 102444   980   5% /lib/rc/init.d
mdev 10240 0 10240   0% /dev
shm1551308 0   1551308   0% /dev/shm
/dev/sda7483917384 251951296 231966088  53% /home

  The secret is that I bindmount /opt, /var, /usr, and /tmp onto the
large reiserfs partition.
##
/dev/sda5   / ext2 noatime,nodiratime,async0 1
/dev/sda7   /home reiserfs noatime,nodiratime,async,notail 0 1
/home/bindmounts/opt/opt  auto bind0 0
/home/bindmounts/var/var  auto bind0 0
/home/bindmounts/usr/usr  auto bind0 0
/home/bindmounts/tmp/tmp  auto bind0 0
/dev/sda6   noneswapsw  0 0
/dev/cdrom  /mnt/cdrom  iso9660 noauto,users,ro 0 0
/dev/cdrom1 /mnt/cdrom1 autonoauto,user,ro  0 0
/dev/sdb1 /mnt/extb   auto  noauto,user,noatime,async   0 0
/dev/sdc1 /mnt/extc   auto  noauto,user,noatime,async   0 0

# glibc 2.2 and above expects tmpfs to be mounted at /dev/shm for 
# POSIX shared memory (shm_open, shm_unlink).
# (tmpfs is a dynamically expandable/shrinkable ramdisk, and will
#  use almost no memory if not populated with files)
shm   /dev/shmtmpfs nodev,nosuid,noexec   0 0
##

-- 
Walter Dnes 

Re: [gentoo-user] Partitioning strategy...?

[Dale]

Pandu Poluan wrote:



On Nov 26, 2011 12:05 AM, "Mark Knecht" > wrote:

>

> I'm exactly as you are WRT to LVM but I admired Dale for giving it a
> shot and I'm sorta feeling like I gotta start learn it just to be part
> of the group... ;-)
>

Hey, not fair! Dale's got a headstart already with multi-partitions :-)


In the meantime, before I have to rebuild, I'm going to learn me some 
LVM goodness...


Rgds,



I have this on mine:

/
/boot
/home
/usr.portage
/var
/var/tmp/portage on tmpfs.  Nice to have 16Gbs of ram.  :-)

sda1Primaryext2[boot]197.41
sda2Primaryswap  [swap]   1003.49
sda3Primaryreiserfs   [root]20003.89
sda5Logical ext3[blank]   5000.98
sda6Logical ext3[portage]   12000.69
sda7Logical reiserfs   [home]  50001.48
sda8Logical ext3[var]  10001.95


Here is one thing to think about on LVM.  If you put /usr on a separate 
partition, you will need the init thingy, thanks to the dev at fedora.  
Yea, lower case f just like lower case w for winders.  Don't make me 
spell it the way I want.  There is a lady on this list.  o_O  Anyway, if 
you are going to do /usr on a separate partition then you may as well 
have LVM.  You are going to have the init thingy anyway.  You may as 
well give the whole bit a try.


LVM has not really been a problem other than me trying to get my 
sequence and commands straight.  If I was going to install again, I 
would likely have it all on LVM except / and /boot.  After all, this is 
sort of the way fedora does it which is what started the init thingy, in 
my opinion anyway.  I think a Gentoo dev, a really big one, needs to 
poke the fedora dev in the eye, right one since most are right eye 
dominant.  Might make his keyboard look funny for a while.  :/


The init thingy, I have tried making one and booting it.  It fails each 
and every time.  I fix one thing, something else breaks.  Google finds 
the same problems but no fixes.  I can't seem to find a howto that works 
for me, including the Gentoo wiki one.  Dang fedora !


So, LVM, works fine just have to learn it.  The init thing, sucks !!

I think the partitioning scheme varies on what you are doing with your 
box tho.  For home use, /boot, /, /home and maybe /var.  You can do 
/usr/portage if fragmentation bothers you.  I have a /data thing that I 
started way back when I was new on Linux and using Mandrake.  I 
really need to move that stuff to my /home directory.  I was a bit green 
at the time.  lol  I put my TV shows, .iso files and other junk on 
there.  That is on LVM since it seems to grow.


Oh, here is a funny one.  Imagine walking up to the computer and seeing 
knotify taking up 14Gbs of ram.  O_O  My rig was using all the ram, some 
cache and slow as leap year.  I kill -9'd that thing.  G!!


Dale

:-)  :-)

--
I am only responsible for what I said ... Not for what you understood or how 
you interpreted my words!

Re: [gentoo-user] Partitioning strategy...?

[Walter Dnes]

On Fri, Nov 25, 2011 at 10:17:07AM -0800, Mark Knecht wrote
> On Fri, Nov 25, 2011 at 9:53 AM, Pandu Poluan  wrote:
> 
> >> /mnt/distfiles ext3/4 (shared via nfs)
> >>
> >
> > Why do you separate the distfiles from the portage tree?
> >
> > Rgds,
> 
> distfiles has a tendency to grow large over the years. IIRC nothing
> cleans it up automatically so having it separate is just a simple
> safety mechanism to not run out of disk space after emerge -fDuN
> @world, etc.

  man eclean
http://gpio.ca/cgi-bin/man/man2html?1+eclean

-- 
Walter Dnes 

Re: [gentoo-user] Error message for several ebuilds when updating world.

[Walter Dnes]

On Fri, Nov 25, 2011 at 01:31:45AM -0600, Dale wrote

> http://forums.gentoo.org/viewtopic-p-6878530.html?sid=924b3c65e5584c4dc81672583d97b85d#6878530
> 
> You need more ram, swap space or maybe drive space for the portage work 
> directory.  Basically, it's running out of space somewhere.

  Thanks for the pointer.  I think I've figured it it out now.  When I
set up my laptop, I copied most of /etc/make.conf verbatim from my
desktop, including the line...
PORTAGE_TMPFS="/dev/shm"

  That works just fine on my desktop with 8 gigs of ram.  Unfortunately,
my laptop only has 4 gigs, so that doesn't work quite so well.  I've
changed that line to...
PORTAGE_TMPFS="/tmp"

  Builds will be slower, no doubt.

-- 
Walter Dnes 

Re: [gentoo-user] Partitioning strategy...?

[Stéphane Guedon]

On Friday 25 November 2011 19:17:07 Mark Knecht wrote:
> On Fri, Nov 25, 2011 at 9:53 AM, Pandu Poluan  wrote:
> 
> 
> >> /mnt/distfiles ext3/4 (shared via nfs)
> > 
> > Why do you separate the distfiles from the portage tree?
> > 
> > Rgds,
> 
> distfiles has a tendency to grow large over the years. IIRC nothing
> cleans it up automatically so having it separate is just a simple
> safety mechanism to not run out of disk space after emerge -fDuN
> @world, etc.
> 
> I do it also.
> 
> - Mark

It is also because the portage tree is reiserfs and distfiles ext4.

Actually, it is like it :

/usr
/portage-> reiserfs | both shared through nfs
/distfiles  -> ext4 |

/usr
/portage/distfiles  | acces on an nfs

When you mount a filesystem B inside an other one A and share the root A 
through nfs, it seems you acces (from nfs clients) to the A and the directory 
under which is mounted B, but not B itself.

Do you understand me ?

-- 
Stéphane Guedon
page web : http://www.22decembre.eu/
carte de visite : http://www.22decembre.eu/downloads/Stephane-Guedon.vcf
clé publique gpg : http://www.22decembre.eu/downloads/Stephane-Guedon.asc


signature.asc
Description: This is a digitally signed message part.

[gentoo-user] Re: Partitioning strategy...?

[Francesco Talamona]

On Friday 25 November 2011, Pandu Poluan wrote:
> On Fri, Nov 25, 2011 at 20:53, Pandu Poluan  wrote:
> > I want to build a Gentoo server box whose structure is
> > highly-partitioned, like this:
> > 
> > /
> > /boot
> > /usr
> > /tmp
> > /usr/portage ==> via NFS
> > /var
> > /var/lib/postgresql
> > /var/tmp
> > /var/log
> > /var/spool
> > 
> > (Not all of them will reside on the same physical disk; I have
> > /dev/sda up to /dev/sdd)
> > 
> > I've been searching high and low for recommended numbers... and
> > there are as many number as search-hits.
> > 
> > So. Care to share your partitioning strategy?
> > 
> > (And while we're at it, am I overdoing the partitioning?)
> 
> I also wonder if /var/tmp can be shared between boxen (via NFS),
> assuming I ensure that no two boxen perform emerge at the same
> time...
> 
> Rgds,


If you want to mount /var/tmp/portage over NFS you can set in 
/etc/make.conf for every machine:

PORTAGE_TMPDIR=/var/tmp/

So you won't have to worry about concurrent emerge across different 
machines.

Disclaimer: this is an untested idea.

HTH
Francesco


-- 
Linux Version 3.1.1-gentoo, Compiled #1 SMP PREEMPT Mon Nov 14 07:03:50 
CET 2011
Two 1GHz AMD Athlon 64 X2 Processors, 4GB RAM, 4021.84 Bogomips Total
aemaeth

Re: [gentoo-user] glibc-2.14.1 upgrade

[Florian Philipp]

Am 25.11.2011 19:11, schrieb Pandu Poluan:
> 
> On Nov 26, 2011 12:05 AM, "微菜"  > wrote:
>>
>> 于 2011年11月24日 15:34, justin 写道:
>> > On 24/11/11 06:27, Paul Hartman wrote:
>> >> Hi,
>> >>
>> >> After emerging glibc-2.14.1 today, pam stopped working, which
>> >> prevented KDE from working and some other things. I got this kind of
>> >> message:
>> >>
>> >> /lib64/libc.so.6: version `GLIBC_2.14' not found (required by
>> >> /lib64/libcrypt.so.1)
>> >>
>> >> There were no @preserved-rebuild and revdep-rebuild found nothing. I
>> >> rebuilt pam and things seem to be working again. Are there any other
>> >> packages I should rebuild before encountering a problem? Or some way
>> >> to detect which need to be rebuilt? Should I re-emerge world against
>> >> my new glibc? :)
>> >>
>> >> Thanks,
>> >> Paul
>> >>
>> >
>> > Hi Paul,
>> >
>> > after an glibc upgrade it is time to do an
>> >
>> > emerge -e system world
>>
>> you must be kidding me . it will take days to complete.
>>
> 
> On my virtualized servers, emerge -e @system @world, a total of
> 170-something packages, takes only about 6-7 hours. Remember, these are
> VMs, which means the 2 vCPUs I assigned them are actually shared with
> other VMs.
> 
> On a non-virtualized system, should be faster. Of course, if you have
> sizable packages, it might take around 24 hours to complete.
> 
> Rgds,
> 

I've used Gentoo since 2006 and never had any reason to emerge -e. I can
also cite posts on this mailing list going back as far as 2007 (just my
local archive, mind you) that say it is unnecessary for glibc updates.
Or, to cite Volker Armin Hemmann's eloquent reply from 08-11-2007: "no" ;-)

Regards,
Florian Philipp



signature.asc
Description: OpenPGP digital signature

[gentoo-user] Re: Partitioning strategy...?

[Francesco Talamona]

On Friday 25 November 2011, Pandu Poluan wrote:
> On Fri, Nov 25, 2011 at 21:35, Mark Knecht  
wrote:
> > On Fri, Nov 25, 2011 at 5:53 AM, Pandu Poluan  
wrote:
> >> I want to build a Gentoo server box whose structure is
> 
> >> highly-partitioned, like this:
> - >8 snip
> 
> > I don't use LVM but I suspect that on this list that would be the
> > #1 recommendation to take care of the numbers question.
> 
> Everytime I read some guide on LVM, my eyes becomes blurry, the room
> starts spinning, and I can hear wolves howling ... :D
> 
> Seriously, LVM looks mighty nice, but it also looks (and is!) mighty
> complex.
> 
> So, I want to start from something simple.
> 
> Aaaanyways, after reviewing my production boxes, I decided to
> implement the following strategy:
> 
> /  == 800 MiB
> /boot == 20 MiB
> /usr == 1800 MiB
> /usr/portage == 2000 MiB
> /var == 4000 MiB
> /var/lib/postgresql == 1000 MiB
> 
> Comments, suggestions, are welcome :)
> 
> Rgds,

IMO /usr/portage and /boot are too small, these are the respective sizes 
in my system:

[root@aemaeth:~]$ du -sh /usr/portage
19G /usr/portage

[root@aemaeth:~]$ du -sh /boot
52M /boot

HTH
Francesco


-- 
Linux Version 3.1.1-gentoo, Compiled #1 SMP PREEMPT Mon Nov 14 07:03:50 
CET 2011
Two 1GHz AMD Athlon 64 X2 Processors, 4GB RAM, 4021.84 Bogomips Total
aemaeth

Re: [gentoo-user] Re: Partitioning strategy...?

[Pandu Poluan]

On Nov 26, 2011 1:05 AM, "James"  wrote:
>
> Pandu Poluan  poluan.info> writes:
>
>
> > Everytime I read some guide on LVM, my eyes becomes blurry, the room
> > starts spinning, and I can hear wolves howling ... :D
> > Seriously, LVM looks mighty nice, but it also looks (and is!) mighty
complex.
>
> I feel your pain
> I too have had trouble sorting out new installs with raid, GPT,
> and LVM.
>
> Have you seen these guides?
>
> http://www.gentoo.org/doc/en/gentoo-x86+raid+lvm2-quickinstall.xml

This one I've read. Okay, glanced. After finding myself starting blankly at
the screen for a couple of minutes, I decided to read it again later :-)

> http://www.rodsbooks.com/gdisk/
>

Haven't seen this one before. Gotta check it out. Thanks!

> > So, I want to start from something simple.
> > Comments, suggestions, are welcome :)
>
> (OK)
>
> Well the problem is multifaceted, imho, with LVM being just
> a singular issue among the mix. Grub is evolving and the old
> grub has troubles with RAID. Add mdadm, disk over 2T, UUID and the
> issues becomes really murky quickly.
>
> What (IMHO) needs to happen, is the community needs to write some
> install guides, based on notes from several installations, that
> allow for various types of installations (with explicit syntax in-line)
> that starts from simple to complex.
>
> If we keep using  the same installation semantics (examples)
> then the only thing that will change is the additional information on
> the installation complexity.  We could use the new gentoo wiki
> for development.  I know much of this is redundant with the handbook
> for installation, which would still be the "master reference" for
> installations, but there would be a multiplicative example base
> to compliment the handbook and more specifically focused to the
> issues of a given installation. There is precedence for this; The
> handbook already has version for different hardware architectures.
>
> So what I'm proposing is that when someone feels motivated, keep notes
> on your particular installation details, and post the notes (as
> a work in progress) to the gentoo wiki. Then the next time someone
> performs an installation, then look at the 'work in progress', use
> the example, edit (add more detail) to the example, and update
> the wiki. Over time these guides, focused on a particularly specific
> example, could be referenced along with the installation handbook,
> as a compliment. Just look at the handbook in section 4, Preparing
> the disk. Woefully antiquated!
>

Actually, I have been keeping notes of all my installs, complete with
remarks to keep reminding my why I did this or that.

More for my own documentation, actually. I've honed my installation
procedure so much that it only bears a passing resemblance to the handbook.

I'll try to distill the notes and see if I can upload it to the wiki.

>
> So I would also break it down into (2) main examples. One with a very
simple
> boot/root/swap scheme and another with many physically separate
partitions,
> such as (Pandu) seeks.  In the second example of many (maximum)
partitions, a
> discussion of the merits, such as why /tmp should be on a separate
> partition and such could be included. In fact, if only these (2) examples
were
> developed, we could removed the parts of the installation
> instructions, such as GPT, or LVM or RAID in order to create
> the other simpler installation instruction guides.  Also how
> you reference the drives (UUID) in the fstab is an
> integral part of the installation landscape, that is changing.
> Not to mention legacy bios and the latest issue of UEFI.
>

I was once somewhat familiar with UUID-based fstab when I was still using
Ubuntu. Too bad I've deleted my last Ubuntu VM a couple of weeks ago. Let's
see if I can still find my installation notes...

Rgds,

Re: [gentoo-user] Partitioning strategy...?

[Mark Knecht]

On Fri, Nov 25, 2011 at 9:53 AM, Pandu Poluan  wrote:

>> /mnt/distfiles ext3/4 (shared via nfs)
>>
>
> Why do you separate the distfiles from the portage tree?
>
> Rgds,

distfiles has a tendency to grow large over the years. IIRC nothing
cleans it up automatically so having it separate is just a simple
safety mechanism to not run out of disk space after emerge -fDuN
@world, etc.

I do it also.

- Mark

Re: [gentoo-user] glibc-2.14.1 upgrade

[Pandu Poluan]

On Nov 26, 2011 12:05 AM, "微菜"  wrote:
>
> 于 2011年11月24日 15:34, justin 写道:
> > On 24/11/11 06:27, Paul Hartman wrote:
> >> Hi,
> >>
> >> After emerging glibc-2.14.1 today, pam stopped working, which
> >> prevented KDE from working and some other things. I got this kind of
> >> message:
> >>
> >> /lib64/libc.so.6: version `GLIBC_2.14' not found (required by
> >> /lib64/libcrypt.so.1)
> >>
> >> There were no @preserved-rebuild and revdep-rebuild found nothing. I
> >> rebuilt pam and things seem to be working again. Are there any other
> >> packages I should rebuild before encountering a problem? Or some way
> >> to detect which need to be rebuilt? Should I re-emerge world against
> >> my new glibc? :)
> >>
> >> Thanks,
> >> Paul
> >>
> >
> > Hi Paul,
> >
> > after an glibc upgrade it is time to do an
> >
> > emerge -e system world
>
> you must be kidding me . it will take days to complete.
>

On my virtualized servers, emerge -e @system @world, a total of
170-something packages, takes only about 6-7 hours. Remember, these are
VMs, which means the 2 vCPUs I assigned them are actually shared with other
VMs.

On a non-virtualized system, should be faster. Of course, if you have
sizable packages, it might take around 24 hours to complete.

Rgds,

[gentoo-user] Re: Partitioning strategy...?

[James]

Pandu Poluan  poluan.info> writes:


> Everytime I read some guide on LVM, my eyes becomes blurry, the room
> starts spinning, and I can hear wolves howling ... :D
> Seriously, LVM looks mighty nice, but it also looks (and is!) mighty complex.

I feel your pain
I too have had trouble sorting out new installs with raid, GPT, 
and LVM.

Have you seen these guides?

http://www.gentoo.org/doc/en/gentoo-x86+raid+lvm2-quickinstall.xml
http://www.rodsbooks.com/gdisk/

> So, I want to start from something simple.
> Comments, suggestions, are welcome :)

(OK)

Well the problem is multifaceted, imho, with LVM being just 
a singular issue among the mix. Grub is evolving and the old
grub has troubles with RAID. Add mdadm, disk over 2T, UUID and the
issues becomes really murky quickly.

What (IMHO) needs to happen, is the community needs to write some 
install guides, based on notes from several installations, that 
allow for various types of installations (with explicit syntax in-line) 
that starts from simple to complex.

If we keep using  the same installation semantics (examples)
then the only thing that will change is the additional information on
the installation complexity.  We could use the new gentoo wiki 
for development.  I know much of this is redundant with the handbook 
for installation, which would still be the "master reference" for 
installations, but there would be a multiplicative example base 
to compliment the handbook and more specifically focused to the
issues of a given installation. There is precedence for this; The
handbook already has version for different hardware architectures.

So what I'm proposing is that when someone feels motivated, keep notes
on your particular installation details, and post the notes (as
a work in progress) to the gentoo wiki. Then the next time someone
performs an installation, then look at the 'work in progress', use
the example, edit (add more detail) to the example, and update
the wiki. Over time these guides, focused on a particularly specific
example, could be referenced along with the installation handbook,
as a compliment. Just look at the handbook in section 4, Preparing
the disk. Woefully antiquated!


So I would also break it down into (2) main examples. One with a very simple
boot/root/swap scheme and another with many physically separate partitions,
such as (Pandu) seeks.  In the second example of many (maximum) partitions, a
discussion of the merits, such as why /tmp should be on a separate
partition and such could be included. In fact, if only these (2) examples were 
developed, we could removed the parts of the installation 
instructions, such as GPT, or LVM or RAID in order to create 
the other simpler installation instruction guides.  Also how 
you reference the drives (UUID) in the fstab is an 
integral part of the installation landscape, that is changing. 
Not to mention legacy bios and the latest issue of UEFI.


Personally, I've made several attempts to install a new work station
with RAID 1 on boot/root/swap, using 2T seagate drives, some time
ago. I did not want to use LVM and grub would not boot. I think 
I'll give that install another whirl and yes, I'll post to the
wiki, once I get it right.

hth,
James

Re: [gentoo-user] Partitioning strategy...?

[Pandu Poluan]

On Nov 26, 2011 12:06 AM, "Stéphane Guedon"  wrote:
>
> On Friday 25 November 2011 14:53:17 Pandu Poluan wrote:
> > I want to build a Gentoo server box whose structure is
> > highly-partitioned, like this:
> >
> > /
> > /boot
> > /usr
> > /tmp
> > /usr/portage ==> via NFS
> > /var
> > /var/lib/postgresql
> > /var/tmp
> > /var/log
> > /var/spool
> >

- >8 snip

>
> I never set /usr separated from / especially on a server : if there's a
bug
> for any reason, nothing works ! (emerge is in /usr, gcc, ssh doesn't
start).
>
> But you are the one who decide !
>

Well, actually that's the reason why I want to separate /usr: I'm going to
mount it ro to prevent something bad happening to the extremely important
files within.

> This is my partition system :
> /   ext3/4
> /home   ext3/4
> /varreiserfs
> /tmptmpfs
> /tmp_portage tmpfs (specifically for emerge, so I can mount or unmount it
when
> large compil start)
> /mnt/portage reiserfs (shared via nfs)
> /mnt/distfiles ext3/4 (shared via nfs)
>

Why do you separate the distfiles from the portage tree?

Rgds,

Re: [gentoo-user] Partitioning strategy...?

[Pandu Poluan]

On Nov 26, 2011 12:05 AM, "Mark Knecht"  wrote:
>
> On Fri, Nov 25, 2011 at 8:12 AM, Pandu Poluan  wrote:
> > On Fri, Nov 25, 2011 at 21:35, Mark Knecht  wrote:
> >> On Fri, Nov 25, 2011 at 5:53 AM, Pandu Poluan 
wrote:
> >>> I want to build a Gentoo server box whose structure is
> >>> highly-partitioned, like this:
> >>>
> >
> > - >8 snip
> >
> >>
> >> I don't use LVM but I suspect that on this list that would be the #1
> >> recommendation to take care of the numbers question.
> >>
> >
> > Everytime I read some guide on LVM, my eyes becomes blurry, the room
> > starts spinning, and I can hear wolves howling ... :D
> >
> > Seriously, LVM looks mighty nice, but it also looks (and is!) mighty
complex.
> >
>
> I'm exactly as you are WRT to LVM but I admired Dale for giving it a
> shot and I'm sorta feeling like I gotta start learn it just to be part
> of the group... ;-)
>

Hey, not fair! Dale's got a headstart already with multi-partitions :-)

> Seriously though, I've done enough RAID (0,1,5 & 6) recently to at
> least feel comfortable setting it up. I'm much more worried about
> whether I'll be able to handle it when it eventually breaks down. None
> the less, with LVM on top of RAID I think I'd get past a lot of
> limitations that I run into, and you may, when I picked a certain size
> and 1 year down the road it turns out my needs changed.
>

Well, if the numbers I've chosen prove to be off the mark, I'll just
tarball everything and rebuild :-P

In the meantime, before I have to rebuild, I'm going to learn me some LVM
goodness...

Rgds,

Re: [gentoo-user] Partitioning strategy...?

[Stéphane Guedon]

On Friday 25 November 2011 14:53:17 Pandu Poluan wrote:
> I want to build a Gentoo server box whose structure is
> highly-partitioned, like this:
> 
> /
> /boot
> /usr
> /tmp
> /usr/portage ==> via NFS
> /var
> /var/lib/postgresql
> /var/tmp
> /var/log
> /var/spool
> 
> (Not all of them will reside on the same physical disk; I have
> /dev/sda up to /dev/sdd)
> 
> I've been searching high and low for recommended numbers... and there
> are as many number as search-hits.
> 
> So. Care to share your partitioning strategy?
> 
> (And while we're at it, am I overdoing the partitioning?)
> 
> Rgds

I never set /usr separated from / especially on a server : if there's a bug 
for any reason, nothing works ! (emerge is in /usr, gcc, ssh doesn't start).

But you are the one who decide !

This is my partition system :
/   ext3/4
/home   ext3/4
/varreiserfs
/tmptmpfs
/tmp_portage tmpfs (specifically for emerge, so I can mount or unmount it when 
large compil start)
/mnt/portage reiserfs (shared via nfs)
/mnt/distfiles ext3/4 (shared via nfs)

I am available for any explanation. For the ones who read french I have 
written a doc on my website concerning my choices.
-- 
Stéphane Guedon
page web : http://www.22decembre.eu/
carte de visite : http://www.22decembre.eu/downloads/Stephane-Guedon.vcf
clé publique gpg : http://www.22decembre.eu/downloads/Stephane-Guedon.asc


signature.asc
Description: This is a digitally signed message part.

Re: [gentoo-user] Re: experience with rsnapshot

[covici]

Harry Putnam  wrote:

> cov...@ccs.covici.com writes:
> 
> > Grant Edwards  wrote:
> >
> >> On 2011-11-24, cov...@ccs.covici.com  wrote:
> >> 
> >> > I am using rdiff-backup which is no longer maintained, but still seems
> >> > to work, but I was thinking to use rsnapshot instead which seems like a
> >> > nice way to do this, but this seems not to have been maintained for a
> >> > while, either, so I was wondering if anyone is using it and how it works
> >> > for you?
> >> 
> >> I set up rsnapshot a few months ago, and so far it seems to be working
> >> fine.  I found the documentation about how to configure the intervals
> >> and schedule the jobs to be a bit confusing, but once the light bulb
> >> went on, it's pretty easy.
> >
> >
> > Thanks, this is what I was wondering about.
> 
> I'll chime in a bit here too.  I've used rsnapshot for actual yrs,
> maybe 3-4.  I've needed the occasional buggered up file from the
> backups and few whole directories over the years.
> 
> It does not claim any baremetal restore capability... unless its been
> added.  I know there is quite a lot of new functionality that I have
> not had occasion to delve into.
> 
> It does not afford a handy slick way of retrieving a backed up file.
> I mean it is left to your own devices... but since the increments are
> dated and in hourly, daily, weekly, monthly [...] groupings, its not
> so hard to find what you need... I'm just saying it is a manual
> process unless you script something. 
> 
> I probably should investigate new features... since the above may be
> outdated information.
> 
> One thing you can be sure of... its highly reliable since it is based
> on a very robust and well tested rsync and a very robust perl.  Also,
> you will be amazed at how many backups you can have and take so very
> little space.
> 
> Of course that last will depend to a good degree how much actual
> change occurs in your data being backed up.
> 
> Further, it lends itself to network activity very well.
> 
> All in all a quite simple to use, highly reliable, network capable,
> very versatile system.
> 

I noticed that there was no real restore, but as you say you can usually
find what you are looking for.  I will probably try on an experimental
basis.

Thanks for all your responses.

-- 
Your life is like a penny.  You're going to lose it.  The question is:
How do
you spend it?

 John Covici
 cov...@ccs.covici.com

Re: [gentoo-user] glibc-2.14.1 upgrade

[微菜]

于 2011年11月24日 15:34, justin 写道:
> On 24/11/11 06:27, Paul Hartman wrote:
>> Hi,
>>
>> After emerging glibc-2.14.1 today, pam stopped working, which
>> prevented KDE from working and some other things. I got this kind of
>> message:
>>
>> /lib64/libc.so.6: version `GLIBC_2.14' not found (required by
>> /lib64/libcrypt.so.1)
>>
>> There were no @preserved-rebuild and revdep-rebuild found nothing. I
>> rebuilt pam and things seem to be working again. Are there any other
>> packages I should rebuild before encountering a problem? Or some way
>> to detect which need to be rebuilt? Should I re-emerge world against
>> my new glibc? :)
>>
>> Thanks,
>> Paul
>>
> 
> Hi Paul,
> 
> after an glibc upgrade it is time to do an
> 
> emerge -e system world

you must be kidding me . it will take days to complete.

> 
> 
> This is what binary distros actually do when they have major bumps in
> their releases.
> 
> Justin
> 


-- 
我是天马博士，对，就是创造了阿童木的那个天马博士



signature.asc
Description: OpenPGP digital signature

Re: [gentoo-user] Partitioning strategy...?

[Mark Knecht]

On Fri, Nov 25, 2011 at 8:12 AM, Pandu Poluan  wrote:
> On Fri, Nov 25, 2011 at 21:35, Mark Knecht  wrote:
>> On Fri, Nov 25, 2011 at 5:53 AM, Pandu Poluan  wrote:
>>> I want to build a Gentoo server box whose structure is
>>> highly-partitioned, like this:
>>>
>
> - >8 snip
>
>>
>> I don't use LVM but I suspect that on this list that would be the #1
>> recommendation to take care of the numbers question.
>>
>
> Everytime I read some guide on LVM, my eyes becomes blurry, the room
> starts spinning, and I can hear wolves howling ... :D
>
> Seriously, LVM looks mighty nice, but it also looks (and is!) mighty complex.
>

I'm exactly as you are WRT to LVM but I admired Dale for giving it a
shot and I'm sorta feeling like I gotta start learn it just to be part
of the group... ;-)

Seriously though, I've done enough RAID (0,1,5 & 6) recently to at
least feel comfortable setting it up. I'm much more worried about
whether I'll be able to handle it when it eventually breaks down. None
the less, with LVM on top of RAID I think I'd get past a lot of
limitations that I run into, and you may, when I picked a certain size
and 1 year down the road it turns out my needs changed.

Still, I glaze over every time... :-)

- Mark

Re: [gentoo-user] Partitioning strategy...?

[Jarry]

On 25-Nov-11 17:12, Pandu Poluan wrote:


/  == 800 MiB
/boot == 20 MiB
/usr == 1800 MiB
/usr/portage == 2000 MiB
/var == 4000 MiB
/var/lib/postgresql == 1000 MiB


I think it is more than wise to put /tmp on separate
partition, and mount it with nodev/nosuid/noexec.

Malware frequently use tmp to download & compile
some bad tools and run them from there, as tmp
is one of not many world-writable directories...

Jarry
--
___
This mailbox accepts e-mails only from selected mailing-lists!
Everything else is considered to be spam and therefore deleted.

Re: [gentoo-user] Partitioning strategy...?

[Pandu Poluan]

On Fri, Nov 25, 2011 at 21:35, Mark Knecht  wrote:
> On Fri, Nov 25, 2011 at 5:53 AM, Pandu Poluan  wrote:
>> I want to build a Gentoo server box whose structure is
>> highly-partitioned, like this:
>>

- >8 snip

>
> I don't use LVM but I suspect that on this list that would be the #1
> recommendation to take care of the numbers question.
>

Everytime I read some guide on LVM, my eyes becomes blurry, the room
starts spinning, and I can hear wolves howling ... :D

Seriously, LVM looks mighty nice, but it also looks (and is!) mighty complex.

So, I want to start from something simple.

Aaaanyways, after reviewing my production boxes, I decided to
implement the following strategy:

/  == 800 MiB
/boot == 20 MiB
/usr == 1800 MiB
/usr/portage == 2000 MiB
/var == 4000 MiB
/var/lib/postgresql == 1000 MiB

Comments, suggestions, are welcome :)

Rgds,
-- 
FdS Pandu E Poluan
~ IT Optimizer ~

 • LOPSA Member #15248
 • Blog : http://pepoluan.tumblr.com
 • Linked-In : http://id.linkedin.com/in/pepoluan

[gentoo-user] Re: experience with rsnapshot

[Harry Putnam]

cov...@ccs.covici.com writes:

> Grant Edwards  wrote:
>
>> On 2011-11-24, cov...@ccs.covici.com  wrote:
>> 
>> > I am using rdiff-backup which is no longer maintained, but still seems
>> > to work, but I was thinking to use rsnapshot instead which seems like a
>> > nice way to do this, but this seems not to have been maintained for a
>> > while, either, so I was wondering if anyone is using it and how it works
>> > for you?
>> 
>> I set up rsnapshot a few months ago, and so far it seems to be working
>> fine.  I found the documentation about how to configure the intervals
>> and schedule the jobs to be a bit confusing, but once the light bulb
>> went on, it's pretty easy.
>
>
> Thanks, this is what I was wondering about.

I'll chime in a bit here too.  I've used rsnapshot for actual yrs,
maybe 3-4.  I've needed the occasional buggered up file from the
backups and few whole directories over the years.

It does not claim any baremetal restore capability... unless its been
added.  I know there is quite a lot of new functionality that I have
not had occasion to delve into.

It does not afford a handy slick way of retrieving a backed up file.
I mean it is left to your own devices... but since the increments are
dated and in hourly, daily, weekly, monthly [...] groupings, its not
so hard to find what you need... I'm just saying it is a manual
process unless you script something. 

I probably should investigate new features... since the above may be
outdated information.

One thing you can be sure of... its highly reliable since it is based
on a very robust and well tested rsync and a very robust perl.  Also,
you will be amazed at how many backups you can have and take so very
little space.

Of course that last will depend to a good degree how much actual
change occurs in your data being backed up.

Further, it lends itself to network activity very well.

All in all a quite simple to use, highly reliable, network capable,
very versatile system.

Re: [gentoo-user] Partitioning strategy...?

[Mark Knecht]

On Fri, Nov 25, 2011 at 5:53 AM, Pandu Poluan  wrote:
> I want to build a Gentoo server box whose structure is
> highly-partitioned, like this:
>
> /
> /boot
> /usr
> /tmp
> /usr/portage ==> via NFS
> /var
> /var/lib/postgresql
> /var/tmp
> /var/log
> /var/spool
>
> (Not all of them will reside on the same physical disk; I have
> /dev/sda up to /dev/sdd)
>
> I've been searching high and low for recommended numbers... and there
> are as many number as search-hits.
>
> So. Care to share your partitioning strategy?
>
> (And while we're at it, am I overdoing the partitioning?)
>
> Rgds,
> --
> FdS Pandu E Poluan
> ~ IT Optimizer ~
>
>  • LOPSA Member #15248
>  • Blog : http://pepoluan.tumblr.com
>  • Linked-In : http://id.linkedin.com/in/pepoluan
>
>

I don't use LVM but I suspect that on this list that would be the #1
recommendation to take care of the numbers question.

- Mark

[gentoo-user] Re: Partitioning strategy...?

[Pandu Poluan]

On Fri, Nov 25, 2011 at 20:53, Pandu Poluan  wrote:
> I want to build a Gentoo server box whose structure is
> highly-partitioned, like this:
>
> /
> /boot
> /usr
> /tmp
> /usr/portage ==> via NFS
> /var
> /var/lib/postgresql
> /var/tmp
> /var/log
> /var/spool
>
> (Not all of them will reside on the same physical disk; I have
> /dev/sda up to /dev/sdd)
>
> I've been searching high and low for recommended numbers... and there
> are as many number as search-hits.
>
> So. Care to share your partitioning strategy?
>
> (And while we're at it, am I overdoing the partitioning?)
>

I also wonder if /var/tmp can be shared between boxen (via NFS),
assuming I ensure that no two boxen perform emerge at the same time...

Rgds,
-- 
FdS Pandu E Poluan
~ IT Optimizer ~

 • LOPSA Member #15248
 • Blog : http://pepoluan.tumblr.com
 • Linked-In : http://id.linkedin.com/in/pepoluan

[gentoo-user] Partitioning strategy...?

[Pandu Poluan]

I want to build a Gentoo server box whose structure is
highly-partitioned, like this:

/
/boot
/usr
/tmp
/usr/portage ==> via NFS
/var
/var/lib/postgresql
/var/tmp
/var/log
/var/spool

(Not all of them will reside on the same physical disk; I have
/dev/sda up to /dev/sdd)

I've been searching high and low for recommended numbers... and there
are as many number as search-hits.

So. Care to share your partitioning strategy?

(And while we're at it, am I overdoing the partitioning?)

Rgds,
-- 
FdS Pandu E Poluan
~ IT Optimizer ~

 • LOPSA Member #15248
 • Blog : http://pepoluan.tumblr.com
 • Linked-In : http://id.linkedin.com/in/pepoluan