[gentoo-user] Make QEMU guest visible to other machines on LAN

[waltdnes]

  Apologies if this is a duplicate/triplicate.  I don't think the first
attempts got through.  Going through my backup provider (dialup) this
time.

  I have QEMU installed on a 64-bit Gentoo machine.  I'm now installing
a 32-bit Gentoo guest.  The "cdrom" (actually the minimal install ISO
file) boots, and dhcpcd hands out IP address 10.0.2.15 and gateway
10.0.2.2.  The install can connect to the outside world via the "links"
browser and it can ssh into the host machine 192.168.123.249 and visa
versa.  But the host ssh's into the guest install session via a port
redirection into itself. ( ssh -p  localhost ) 

  For various reasons, I need another physical machine on my small home
LAN to be able to talk directly to the 32-bit guest.  I've read the
"Network setup" at http://www.linux-kvm.org/page/KvmOnGentooIs the
/etc/conf.d/net being reffered to, the one on the host or on the guest?
The webpage doesn't say explicitly.

  I also don't understand how other machines will be able to
differentiate between the host and the guest.

-- 
Walter Dnes 
I don't run "desktop environments"; I run useful applications

Re: [gentoo-user] no network "eth0" after upgrade.

[Alan Mackenzie]

Hello, Thelma.

On Sat, Dec 19, 2015 at 12:10:58PM -0700, the...@sys-concept.com wrote:
> I just upgraded one of my systems and upon boot there is no network:

> /etc/init.d/net.eth0 start
> * Mounting local filesystems ...
> mount: mount point /proc/bus/usb does not exist
> * Some local filesystem failed to start
> ERROR: localmount failed to start
> ERROR: cannot start net.eth0 as localmount would not start

> rc-status sysinit
> sysfs   [  started  ]
> dmesg   [  started  ]
> devfs  [  started  ]
> tmpfiles.dev  [  started  ]
> udev   [  started


> I can not even ssh to the system as network is not working.
> What to check next?

Have you got CONFIG_FHANDLE set in your kernel config?  This almost
tripped me up when I synched a day or two ago.  The new udev ebuild
output a message saying this was needed.  It wasn't wrong, either.

CONFIG_FHANDLE can be found in General Setup, under the prompt "open by
fhandle syscalls".

> -- 
> Thelma

-- 
Alan Mackenzie (Nuremberg, Germany).

Re: [gentoo-user] anyone tried amdgpu (kernel module)

[Alexander Kapshuk]

On Sat, Dec 19, 2015 at 5:43 PM, Tsukasa Mcp_Reznor 
wrote:

> From: alexander.kaps...@gmail.com
> Date: Sat, 19 Dec 2015 09:35:28 +0200
> Subject: Re: [gentoo-user] anyone tried amdgpu (kernel module)
> To: gentoo-user@lists.gentoo.org
>
> On Fri, Dec 18, 2015 at 11:07 PM, Tsukasa Mcp_Reznor <
> mcp_rez...@hotmail.com> wrote:
>
> I have a Bonaire gpu, which has legacy support using the amdgpu kernel
> module.  I currently use the Radeon dri module with radeonsi mesa drivers
> and am quite happy.  But gentoo being gentoo I thought I'd give the amdgpu
> a go for the fun of it.
>
> Tried a few variations and keep coming up with a black screen on boot and
> it's hard locked.  I'm using the same Firmware includes that the radeon
> driver requires, and from reading it looks like that's fine with amdgpu, so
> I'm not sure what else could be the issue.
>
> Has anyone here tried and had success with it?  I've tried Kernels 4.2
> 4.2.4 4.3 and 4.3.2.  So I believe I'm missing something simple and it's
> not a kernel bug.
>
>
> Did you consult the wiki article shown below when configuring your system
> to use admgpu?
>
> https://wiki.gentoo.org/wiki/Amdgpu
>
>
>
>
>
> Yes I have.  The CIK parts kernel option is enabled.  the only thing I
> find a bit odd, is in the firmware include list is radeon/bonaire_sdma1.bin
> which doesn't exist on my system using sys-kernel/linux-firmware-20150812.
> So I'm assuming it's an error on the wiki page.
>

Have you tried this firmware package instead, sys-firmware/amdgpu-ucode?

Do you have the firmware included in your kernel config file?

Device Drivers  --->
Generic Driver Options  --->
-*- Userspace firmware loading support
[*] Include in-kernel firmware blobs in kernel binary
(amdgpu/.bin radeon/.bin)
(/lib/firmware) Firmware blobs root directory

Re: [gentoo-user] no network "eth0" after upgrade.

[thelma]

On 12/19/2015 12:10 PM, the...@sys-concept.com wrote:
> I just upgraded one of my systems and upon boot there is no network:
> 
> /etc/init.d/net.eth0 start
> * Mounting local filesystems ...
> mount: mount point /proc/bus/usb does not exist
> * Some local filesystem failed to start
> ERROR: localmount failed to start
> ERROR: cannot start net.eth0 as localmount would not start
> 
> rc-status sysinit
> sysfs   [  started  ]
> dmesg   [  started  ]
> devfs  [  started  ]
> tmpfiles.dev  [  started  ]
> udev   [  started
> 
> 
> I can not even ssh to the system as network is not working.
> What to check next?
> 
It seems I'm not the only one:
https://forums.gentoo.org/viewtopic-t-1034770-highlight-localmount.html

"The problem was openrc-0.18.4. When I downgraded to openrc-0.16.4 the
problem went away."

Now, I can not downgrade without eth0 working.
Do I need to boot strap and downgrade or is there is easier solution?

--
Thelma

RE: [gentoo-user] (Solved) anyone tried amdgpu (kernel module)

[Tsukasa Mcp_Reznor]

From: mcp_rez...@hotmail.com
To: gentoo-user@lists.gentoo.org
Subject: RE: [gentoo-user] anyone tried amdgpu (kernel module)
Date: Sat, 19 Dec 2015 11:17:40 -0500






From: alexander.kaps...@gmail.com
Date: Sat, 19 Dec 2015 18:13:03 +0200
Subject: Re: [gentoo-user] anyone tried amdgpu (kernel module)
To: gentoo-user@lists.gentoo.org

On Sat, Dec 19, 2015 at 5:43 PM, Tsukasa Mcp_Reznor  
wrote:



From: alexander.kaps...@gmail.com
Date: Sat, 19 Dec 2015 09:35:28 +0200
Subject: Re: [gentoo-user] anyone tried amdgpu (kernel module)
To: gentoo-user@lists.gentoo.org

On Fri, Dec 18, 2015 at 11:07 PM, Tsukasa Mcp_Reznor  
wrote:



I have a Bonaire gpu, which has legacy support using the amdgpu kernel module.  
I currently use the Radeon dri module with radeonsi mesa drivers and am quite 
happy.  But gentoo being gentoo I thought I'd give the amdgpu a go for the fun 
of it.
Tried a few variations and keep coming up with a black screen on boot and it's 
hard locked.  I'm using the same Firmware includes that the radeon driver 
requires, and from reading it looks like that's fine with amdgpu, so I'm not 
sure what else could be the issue.
Has anyone here tried and had success with it?  I've tried Kernels 4.2 4.2.4 
4.3 and 4.3.2.  So I believe I'm missing something simple and it's not a kernel 
bug. 

Did you consult the wiki article shown below when configuring your system to 
use admgpu?

https://wiki.gentoo.org/wiki/Amdgpu





Yes I have.  The CIK parts kernel option is enabled.  the only thing I find a 
bit odd, is in the firmware include list is radeon/bonaire_sdma1.bin which 
doesn't exist on my system using sys-kernel/linux-firmware-20150812.  So I'm 
assuming it's an error on the wiki page.

Have you tried this firmware package instead, sys-firmware/amdgpu-ucode?

Do you have the firmware included in your kernel config file?
Device Drivers  --->
Generic Driver Options  --->
-*- Userspace firmware loading support
[*] Include in-kernel firmware blobs in kernel binary
(amdgpu/.bin radeon/.bin)
(/lib/firmware) Firmware blobs root directory

Yes I have the firmware includes, I'll try the amdgpu-ucode and report back.
-
Mission successful!  Thanks everyone, it appears the missing sdma1.bin from 
linux-firmware is contained in amdgpu-ucode and after switching everything is 
running fine.


  

Re: [gentoo-user] no network "eth0" after upgrade.

[thelma]

On 12/19/2015 02:04 PM, John Runyon wrote:
> Not proc, but you should add nofail to scanner.
> 
> John Runyon
> Sent from my phone
> 
> On Dec 19, 2015 2:59 PM, the...@sys-concept.com wrote:
>>
>> On 12/19/2015 12:57 PM, Alexander Kapshuk wrote: 
>> [snip] 
>>
>
 It seems I'm not the only one: 
 https://forums.gentoo.org/viewtopic-t-1034770-highlight-localmount.html 

 "The problem was openrc-0.18.4. When I downgraded to openrc-0.16.4 the 
 problem went away." 

 Now, I can not downgrade without eth0 working. 
 Do I need to boot strap and downgrade or is there is easier solution? 

 -- 
 Thelma 



>>> Here is a news item that explains the situation. 
>>>
>>> 2015-10-07-openrc-0-18-localmount-and-netmount-changes 
>>>Title OpenRC-0.18 localmount and netmount changes 
>>>AuthorWilliam Hubbs  
>>>Posted2015-10-07 
>>>Revision  1 
>>>
>>> The behaviour of localmount and netmount is changing on Linux systems. 
>>> In the past, these services always started successfully. However, now they 
>>> will fail if a file system they attempt to mount cannot be mounted. 
>>>
>>> If you have file systems listed in fstab which should not be mounted at 
>>> boot time, make sure to add noauto to the mount options. If you have 
>>> file systems that you want to attempt to mount at boot time but failure 
>>> should be allowed, add nofail to the mount options for these file 
>>> systems in fstab. 
>>>
>>
>> This is my fstab: 
>> /dev/hda1 /boot ext2 noauto,noatime 1 1 
>> /dev/hda3 / ext3 noatime 0 1 
>> /dev/hda2 none swap sw 0 0 
>> /dev/hda4 /home ext3 noatime   0 1 
>>
>> /dev/hdd /mnt/cdrom auto noauto,ro,users 0 0 
>> /dev/hdd/mnt/dvdr   auto noauto,users
>> 0 0 
>>
>> # Scanner 
>> none   /proc/bus/usb   usbfs   defaults,devmode=0666   0 0 
>>
>> none /proc proc defaults 0 0 
>>
>> Does it mean I should add: "nofail" to Scanner and "/proc" line? 

OK, I've added "nofail" to Scanner, and still no network after boot.

--
Thelma

[gentoo-user] Re: Hard drive noise

[James]

Dale  gmail.com> writes:


> I finally got a new 3TB hard drive.
> Thoughts? 
> Dale

Lots of good information posted so far. 2 things I think are missed.
Recently, Thailand had some very bad flooding. TV showed clips
of the manufacturing districts Flooded by more than 20 feet. Many
HD manufacturing plants are in that area. Check, if you can the
location of the drive's manufacture. But those folks are very 
smart and probably partially disassembled the drives, shipped them
elsewhere, dried them out and re-assembled the parts


Second and most critical:: Temperature. Monitor it and try to minimized 
thermal cycles on HD (all HD). Sure they will take it, minute differences 
in the "coefficient of expansion" of metals, allowances is what primarily
gives vibration and opportunity to be destructive after many thermal cycles
loosen things up a bit. There are numerous tools to monitor and log the
temps of drives.. Crappy, or loose (specification) drives will fail
faster that better quality drives (statistically). Over ventilation of the
chassis or housing containing the drives is always prudent. In a 4 HD bay
area, mount no more than (2) HD for best results.


hth,
James

[gentoo-user] Re: anyone tried amdgpu (kernel module)

[James]

Tsukasa Mcp_Reznor  hotmail.com> writes:


> I have a Bonaire gpu, which has legacy support using the amdgpu 
> kernel module.  


I'd be most curios to know if anyone has any version of the radeon Fury
(Fury X2) that uses the HBM  running on Gentoo, regardless of configuration.

Multiple cards in the same system?


James

Re: [gentoo-user] (Solved) anyone tried amdgpu (kernel module)

[Alexander Kapshuk]

> Have you tried this firmware package instead, sys-firmware/amdgpu-ucode?
>
> Do you have the firmware included in your kernel config file?
>
> Device Drivers  --->
> Generic Driver Options  --->
> -*- Userspace firmware loading support
> [*] Include in-kernel firmware blobs in kernel binary
> (amdgpu/.bin radeon/.bin)
> (/lib/firmware) Firmware blobs root directory
>
>
>
> Yes I have the firmware includes, I'll try the amdgpu-ucode and report back.
>
>
> -
>
>
> Mission successful!  Thanks everyone, it appears the missing sdma1.bin from 
> linux-firmware is contained in amdgpu-ucode and after switching everything is 
> running fine.
>
>
>
>
Good to hear.

Thanks for letting us know.

Re: [gentoo-user] no network "eth0" after upgrade.

[thelma]

On 12/19/2015 01:13 PM, Alan Mackenzie wrote:
> Hello, Thelma.
> 
> On Sat, Dec 19, 2015 at 12:10:58PM -0700, the...@sys-concept.com wrote:
>> I just upgraded one of my systems and upon boot there is no network:
> 
>> /etc/init.d/net.eth0 start
>> * Mounting local filesystems ...
>> mount: mount point /proc/bus/usb does not exist
>> * Some local filesystem failed to start
>> ERROR: localmount failed to start
>> ERROR: cannot start net.eth0 as localmount would not start
> 
>> rc-status sysinit
>> sysfs   [  started  ]
>> dmesg   [  started  ]
>> devfs  [  started  ]
>> tmpfiles.dev  [  started  ]
>> udev   [  started
> 
> 
>> I can not even ssh to the system as network is not working.
>> What to check next?
> 
> Have you got CONFIG_FHANDLE set in your kernel config?  This almost
> tripped me up when I synched a day or two ago.  The new udev ebuild
> output a message saying this was needed.  It wasn't wrong, either.
> 
> CONFIG_FHANDLE can be found in General Setup, under the prompt "open by
> fhandle syscalls".


I had CONFIG_FHANDLE=y in kernel .config

--
Thelma

Re: [gentoo-user] no network "eth0" after upgrade.

[John Runyon]

Not proc, but you should add nofail to scanner.

John Runyon
Sent from my phone

On Dec 19, 2015 2:59 PM, the...@sys-concept.com wrote:
>
> On 12/19/2015 12:57 PM, Alexander Kapshuk wrote: 
> [snip] 
>
> >>> 
> >> It seems I'm not the only one: 
> >> https://forums.gentoo.org/viewtopic-t-1034770-highlight-localmount.html 
> >> 
> >> "The problem was openrc-0.18.4. When I downgraded to openrc-0.16.4 the 
> >> problem went away." 
> >> 
> >> Now, I can not downgrade without eth0 working. 
> >> Do I need to boot strap and downgrade or is there is easier solution? 
> >> 
> >> -- 
> >> Thelma 
> >> 
> >> 
> >> 
> > Here is a news item that explains the situation. 
> > 
> > 2015-10-07-openrc-0-18-localmount-and-netmount-changes 
> >   Title OpenRC-0.18 localmount and netmount changes 
> >   Author    William Hubbs  
> >   Posted    2015-10-07 
> >   Revision  1 
> > 
> > The behaviour of localmount and netmount is changing on Linux systems. 
> > In the past, these services always started successfully. However, now they 
> > will fail if a file system they attempt to mount cannot be mounted. 
> > 
> > If you have file systems listed in fstab which should not be mounted at 
> > boot time, make sure to add noauto to the mount options. If you have 
> > file systems that you want to attempt to mount at boot time but failure 
> > should be allowed, add nofail to the mount options for these file 
> > systems in fstab. 
> > 
>
> This is my fstab: 
> /dev/hda1 /boot ext2 noauto,noatime 1 1 
> /dev/hda3 / ext3 noatime 0 1 
> /dev/hda2 none swap sw 0 0 
> /dev/hda4 /home ext3 noatime   0 1 
>
> /dev/hdd /mnt/cdrom auto noauto,ro,users 0 0 
> /dev/hdd    /mnt/dvdr   auto noauto,users    
> 0 0 
>
> # Scanner 
> none   /proc/bus/usb   usbfs   defaults,devmode=0666   0 0 
>
> none /proc proc defaults 0 0 
>
> Does it mean I should add: "nofail" to Scanner and "/proc" line? 
>
> -- 
> Thelma 
>

Re: [gentoo-user] Heads up for those who use grub2

[Bertram Scharpf]

Hi,

On Saturday, 19. Dec 2015, 17:57:33 +, Mick wrote:
> 
> http://hmarco.org/bugs/CVE-2015-8370-Grub2-authentication-bypass.html

Decrementing an unsigned int without checking it for zero is
a real boner. Where things like that happen, you will find
more flaws.

Bertram

-- 
Bertram Scharpf
Stuttgart, Deutschland/Germany
http://www.bertram-scharpf.de

[gentoo-user] Heads up for those who use grub2

[Mick]

http://hmarco.org/bugs/CVE-2015-8370-Grub2-authentication-bypass.html

-- 
Regards,
Mick


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

Re: [gentoo-user] no network "eth0" after upgrade.

[Alexander Kapshuk]

On Sat, Dec 19, 2015 at 9:57 PM, Alexander Kapshuk <
alexander.kaps...@gmail.com> wrote:

> On Sat, Dec 19, 2015 at 9:24 PM,  wrote:
>
>>
>> On 12/19/2015 12:10 PM, the...@sys-concept.com wrote:
>> > I just upgraded one of my systems and upon boot there is no network:
>> >
>> > /etc/init.d/net.eth0 start
>> > * Mounting local filesystems ...
>> > mount: mount point /proc/bus/usb does not exist
>> > * Some local filesystem failed to start
>> > ERROR: localmount failed to start
>> > ERROR: cannot start net.eth0 as localmount would not start
>> >
>> > rc-status sysinit
>> > sysfs   [  started  ]
>> > dmesg   [  started  ]
>> > devfs  [  started  ]
>> > tmpfiles.dev  [  started  ]
>> > udev   [  started
>> >
>> >
>> > I can not even ssh to the system as network is not working.
>> > What to check next?
>> >
>> It seems I'm not the only one:
>> https://forums.gentoo.org/viewtopic-t-1034770-highlight-localmount.html
>>
>> "The problem was openrc-0.18.4. When I downgraded to openrc-0.16.4 the
>> problem went away."
>>
>> Now, I can not downgrade without eth0 working.
>> Do I need to boot strap and downgrade or is there is easier solution?
>>
>> --
>> Thelma
>>
>>
>>
> Here is a news item that explains the situation.
>
> 2015-10-07-openrc-0-18-localmount-and-netmount-changes
>   Title OpenRC-0.18 localmount and netmount changes
>   AuthorWilliam Hubbs 
>   Posted2015-10-07
>   Revision  1
>
> The behaviour of localmount and netmount is changing on Linux systems.
> In the past, these services always started successfully. However, now they
> will fail if a file system they attempt to mount cannot be mounted.
>
> If you have file systems listed in fstab which should not be mounted at
> boot time, make sure to add noauto to the mount options. If you have
> file systems that you want to attempt to mount at boot time but failure
> should be allowed, add nofail to the mount options for these file
> systems in fstab.
>
>
>
As a follow up, here is a postinst message generated for openrc that might
come in handy as well:
>>> Messages generated by process 3631 on 2015-12-03 16:53:14 EET for
package sys-apps/openrc-0.18.4:

WARN: postinst
In this version of OpenRC, the loopback interface no longer
satisfies the net virtual.
If you have services now which do not start because of this,
They can be fixed by adding rc_need="!net"
to the /etc/conf.d/ file.
You should also file a bug against the service asking that
need net be dropped from the dependencies.
The bug you file should block the following tracker:
https://bugs.gentoo.org/show_bug.cgi?id=439092

Bug https://bugs.gentoo.org/show_bug.cgi?id=427996 was not
fixed correctly in earlier versions of OpenRC.
The correct fix is implemented in this version, but that
means netmount needs to be added to the default runlevel if
you are using nfs file systems.

Re: [gentoo-user] no network "eth0" after upgrade.

[Dale]

the...@sys-concept.com wrote:
> On 12/19/2015 02:04 PM, John Runyon wrote:
>> Not proc, but you should add nofail to scanner.
>>
>> John Runyon
>> Sent from my phone
>>
>> On Dec 19, 2015 2:59 PM, the...@sys-concept.com wrote:
>>> On 12/19/2015 12:57 PM, Alexander Kapshuk wrote: 
>>> [snip] 
>>>
> It seems I'm not the only one: 
> https://forums.gentoo.org/viewtopic-t-1034770-highlight-localmount.html 
>
> "The problem was openrc-0.18.4. When I downgraded to openrc-0.16.4 the 
> problem went away." 
>
> Now, I can not downgrade without eth0 working. 
> Do I need to boot strap and downgrade or is there is easier solution? 
>
> -- 
> Thelma 
>
>
>
 Here is a news item that explains the situation. 

 2015-10-07-openrc-0-18-localmount-and-netmount-changes 
Title OpenRC-0.18 localmount and netmount changes 
AuthorWilliam Hubbs  
Posted2015-10-07 
Revision  1 

 The behaviour of localmount and netmount is changing on Linux systems. 
 In the past, these services always started successfully. However, now they 
 will fail if a file system they attempt to mount cannot be mounted. 

 If you have file systems listed in fstab which should not be mounted at 
 boot time, make sure to add noauto to the mount options. If you have 
 file systems that you want to attempt to mount at boot time but failure 
 should be allowed, add nofail to the mount options for these file 
 systems in fstab. 

>>> This is my fstab: 
>>> /dev/hda1 /boot ext2 noauto,noatime 1 1 
>>> /dev/hda3 / ext3 noatime 0 1 
>>> /dev/hda2 none swap sw 0 0 
>>> /dev/hda4 /home ext3 noatime   0 1 
>>>
>>> /dev/hdd /mnt/cdrom auto noauto,ro,users 0 0 
>>> /dev/hdd/mnt/dvdr   auto noauto,users   
>>>  0 0 
>>>
>>> # Scanner 
>>> none   /proc/bus/usb   usbfs   defaults,devmode=0666   0 0 
>>>
>>> none /proc proc defaults 0 0 
>>>
>>> Does it mean I should add: "nofail" to Scanner and "/proc" line? 
> OK, I've added "nofail" to Scanner, and still no network after boot.
>
> --
> Thelma
>
>


I'd try adding nofail to all except / and /home.  Then remove one at a
time until it fails again and that should be the one that is messing
with you. 

Dale

:-)  :-) 

Re: [gentoo-user] Re: Heads up for those who use grub2

[Rich Freeman]

On Sat, Dec 19, 2015 at 4:06 PM, Grant Edwards
 wrote:
> On 2015-12-19, Mick  wrote:
>
>> http://hmarco.org/bugs/CVE-2015-8370-Grub2-authentication-bypass.html
>
> If somebody can touch your computer while it's booting, the game's
> over anyway...
>

Actually, not necessarily, though there is still room to go.

With a TPM-backed full disk encryption scheme you can basically
prevent most attacks based on physical control.  If you were to go a
step further and secure RAM and bus IO (we're not quite there yet) you
could probably make almost any hardware attack completely impractical.
If you have TPM-backed encryption and you assume the software itself
is secure then to attack it you're going to have to actually intercept
data off the bus, or from RAM.  You certainly can't just install some
rootkit by booting from alternate media, or remove the drives and
attack them from another device you control.  That is, unless you
defeat the TPM, which is certainly within the realm of the laws of
physics, but in practice everything about a TPM's design is intended
to prevent that attack.

-- 
Rich

RE: [gentoo-user] anyone tried amdgpu (kernel module)

[Tsukasa Mcp_Reznor]

From: alexander.kaps...@gmail.com
Date: Sat, 19 Dec 2015 09:35:28 +0200
Subject: Re: [gentoo-user] anyone tried amdgpu (kernel module)
To: gentoo-user@lists.gentoo.org

On Fri, Dec 18, 2015 at 11:07 PM, Tsukasa Mcp_Reznor  
wrote:



I have a Bonaire gpu, which has legacy support using the amdgpu kernel module.  
I currently use the Radeon dri module with radeonsi mesa drivers and am quite 
happy.  But gentoo being gentoo I thought I'd give the amdgpu a go for the fun 
of it.
Tried a few variations and keep coming up with a black screen on boot and it's 
hard locked.  I'm using the same Firmware includes that the radeon driver 
requires, and from reading it looks like that's fine with amdgpu, so I'm not 
sure what else could be the issue.
Has anyone here tried and had success with it?  I've tried Kernels 4.2 4.2.4 
4.3 and 4.3.2.  So I believe I'm missing something simple and it's not a kernel 
bug. 

Did you consult the wiki article shown below when configuring your system to 
use admgpu?

https://wiki.gentoo.org/wiki/Amdgpu





Yes I have.  The CIK parts kernel option is enabled.  the only thing I find a 
bit odd, is in the firmware include list is radeon/bonaire_sdma1.bin which 
doesn't exist on my system using sys-kernel/linux-firmware-20150812.  So I'm 
assuming it's an error on the wiki page.

[gentoo-user] no network "eth0" after upgrade.

[thelma]

I just upgraded one of my systems and upon boot there is no network:

/etc/init.d/net.eth0 start
* Mounting local filesystems ...
mount: mount point /proc/bus/usb does not exist
* Some local filesystem failed to start
ERROR: localmount failed to start
ERROR: cannot start net.eth0 as localmount would not start

rc-status sysinit
sysfs   [  started  ]
dmesg   [  started  ]
devfs  [  started  ]
tmpfiles.dev  [  started  ]
udev   [  started


I can not even ssh to the system as network is not working.
What to check next?

-- 
Thelma

Re: [gentoo-user] no network "eth0" after upgrade.

[Alexander Kapshuk]

On Sat, Dec 19, 2015 at 9:24 PM,  wrote:

>
> On 12/19/2015 12:10 PM, the...@sys-concept.com wrote:
> > I just upgraded one of my systems and upon boot there is no network:
> >
> > /etc/init.d/net.eth0 start
> > * Mounting local filesystems ...
> > mount: mount point /proc/bus/usb does not exist
> > * Some local filesystem failed to start
> > ERROR: localmount failed to start
> > ERROR: cannot start net.eth0 as localmount would not start
> >
> > rc-status sysinit
> > sysfs   [  started  ]
> > dmesg   [  started  ]
> > devfs  [  started  ]
> > tmpfiles.dev  [  started  ]
> > udev   [  started
> >
> >
> > I can not even ssh to the system as network is not working.
> > What to check next?
> >
> It seems I'm not the only one:
> https://forums.gentoo.org/viewtopic-t-1034770-highlight-localmount.html
>
> "The problem was openrc-0.18.4. When I downgraded to openrc-0.16.4 the
> problem went away."
>
> Now, I can not downgrade without eth0 working.
> Do I need to boot strap and downgrade or is there is easier solution?
>
> --
> Thelma
>
>
>
Here is a news item that explains the situation.

2015-10-07-openrc-0-18-localmount-and-netmount-changes
  Title OpenRC-0.18 localmount and netmount changes
  AuthorWilliam Hubbs 
  Posted2015-10-07
  Revision  1

The behaviour of localmount and netmount is changing on Linux systems.
In the past, these services always started successfully. However, now they
will fail if a file system they attempt to mount cannot be mounted.

If you have file systems listed in fstab which should not be mounted at
boot time, make sure to add noauto to the mount options. If you have
file systems that you want to attempt to mount at boot time but failure
should be allowed, add nofail to the mount options for these file
systems in fstab.

Re: [gentoo-user] no network "eth0" after upgrade.

[thelma]

On 12/19/2015 12:57 PM, Alexander Kapshuk wrote:
[snip]

>>>
>> It seems I'm not the only one:
>> https://forums.gentoo.org/viewtopic-t-1034770-highlight-localmount.html
>>
>> "The problem was openrc-0.18.4. When I downgraded to openrc-0.16.4 the
>> problem went away."
>>
>> Now, I can not downgrade without eth0 working.
>> Do I need to boot strap and downgrade or is there is easier solution?
>>
>> --
>> Thelma
>>
>>
>>
> Here is a news item that explains the situation.
> 
> 2015-10-07-openrc-0-18-localmount-and-netmount-changes
>   Title OpenRC-0.18 localmount and netmount changes
>   AuthorWilliam Hubbs 
>   Posted2015-10-07
>   Revision  1
> 
> The behaviour of localmount and netmount is changing on Linux systems.
> In the past, these services always started successfully. However, now they
> will fail if a file system they attempt to mount cannot be mounted.
> 
> If you have file systems listed in fstab which should not be mounted at
> boot time, make sure to add noauto to the mount options. If you have
> file systems that you want to attempt to mount at boot time but failure
> should be allowed, add nofail to the mount options for these file
> systems in fstab.
> 

This is my fstab:
/dev/hda1   /boot   ext2noauto,noatime  
1 1
/dev/hda3   /   ext3noatime 
0 1
/dev/hda2   noneswapsw  
0 0
/dev/hda4   /home   ext3noatime 
0 1

/dev/hdd/mnt/cdrom  autonoauto,ro,users 
0 0
/dev/hdd/mnt/dvdr   auto noauto,users0 0

# Scanner
none   /proc/bus/usb   usbfs   defaults,devmode=0666   0 0

none/proc   procdefaults
0 0

Does it mean I should add: "nofail" to Scanner and "/proc" line?

--
Thelma

RE: [gentoo-user] anyone tried amdgpu (kernel module)

[Tsukasa Mcp_Reznor]

From: alexander.kaps...@gmail.com
Date: Sat, 19 Dec 2015 18:13:03 +0200
Subject: Re: [gentoo-user] anyone tried amdgpu (kernel module)
To: gentoo-user@lists.gentoo.org

On Sat, Dec 19, 2015 at 5:43 PM, Tsukasa Mcp_Reznor  
wrote:



From: alexander.kaps...@gmail.com
Date: Sat, 19 Dec 2015 09:35:28 +0200
Subject: Re: [gentoo-user] anyone tried amdgpu (kernel module)
To: gentoo-user@lists.gentoo.org

On Fri, Dec 18, 2015 at 11:07 PM, Tsukasa Mcp_Reznor  
wrote:



I have a Bonaire gpu, which has legacy support using the amdgpu kernel module.  
I currently use the Radeon dri module with radeonsi mesa drivers and am quite 
happy.  But gentoo being gentoo I thought I'd give the amdgpu a go for the fun 
of it.
Tried a few variations and keep coming up with a black screen on boot and it's 
hard locked.  I'm using the same Firmware includes that the radeon driver 
requires, and from reading it looks like that's fine with amdgpu, so I'm not 
sure what else could be the issue.
Has anyone here tried and had success with it?  I've tried Kernels 4.2 4.2.4 
4.3 and 4.3.2.  So I believe I'm missing something simple and it's not a kernel 
bug. 

Did you consult the wiki article shown below when configuring your system to 
use admgpu?

https://wiki.gentoo.org/wiki/Amdgpu





Yes I have.  The CIK parts kernel option is enabled.  the only thing I find a 
bit odd, is in the firmware include list is radeon/bonaire_sdma1.bin which 
doesn't exist on my system using sys-kernel/linux-firmware-20150812.  So I'm 
assuming it's an error on the wiki page.

Have you tried this firmware package instead, sys-firmware/amdgpu-ucode?

Do you have the firmware included in your kernel config file?
Device Drivers  --->
Generic Driver Options  --->
-*- Userspace firmware loading support
[*] Include in-kernel firmware blobs in kernel binary
(amdgpu/.bin radeon/.bin)
(/lib/firmware) Firmware blobs root directory

Yes I have the firmware includes, I'll try the amdgpu-ucode and report back.
  

Re: [gentoo-user] Make QEMU guest visible to other machines on LAN

[Simon Thelen]

On 15-12-19 at 14:21, waltd...@waltdnes.org wrote:
>   Apologies if this is a duplicate/triplicate.  I don't think the first
> attempts got through.  Going through my backup provider (dialup) this
> time.
[..]
>   For various reasons, I need another physical machine on my small home
> LAN to be able to talk directly to the 32-bit guest.  I've read the
> "Network setup" at http://www.linux-kvm.org/page/KvmOnGentooIs the
> /etc/conf.d/net being reffered to, the one on the host or on the guest?
> The webpage doesn't say explicitly.
host
> 
>   I also don't understand how other machines will be able to
> differentiate between the host and the guest.
The idea is to create a bridge device, and then bridge the guests
tun/tap device with the bridge and your ethernet device. This way they
all appear transparently on the network.

The net config on that website should be fine for the host, and then
pass these command options to qemu
-device virtio-net-pci,netdev=hn0,mac="52:54:00:12:34:56" -netdev 
bridge,id=hn0,br=brkvm

You can change the mac address to your liking.

-- 
Simon Thelen

[gentoo-user] Consequences of updating the system on December 13

[gevisz]

After a system update on December 13, 2015, I have found out
that gnome-disks command does not start any more. When trying,
it reports the following:

 $ gnome-disks

(gnome-disks:4828): GLib-GObject-WARNING **: invalid (NULL) pointer instance

(gnome-disks:4828): GLib-GObject-CRITICAL **:
g_signal_handlers_disconnect_matched: assertion 'G_TYPE_CHECK_INSTANCE
(instance)' failed

(gnome-disks:4828): GLib-GObject-WARNING **: invalid (NULL) pointer instance

(gnome-disks:4828): GLib-GObject-CRITICAL **:
g_signal_handlers_disconnect_matched: assertion 'G_TYPE_CHECK_INSTANCE
(instance)' failed

(gnome-disks:4828): GLib-GObject-WARNING **: invalid (NULL) pointer instance

(gnome-disks:4828): GLib-GObject-CRITICAL **:
g_signal_handlers_disconnect_matched: assertion 'G_TYPE_CHECK_INSTANCE
(instance)' failed

(gnome-disks:4828): GLib-GObject-WARNING **: invalid (NULL) pointer instance

(gnome-disks:4828): GLib-GObject-CRITICAL **:
g_signal_handlers_disconnect_matched: assertion 'G_TYPE_CHECK_INSTANCE
(instance)' failed

(gnome-disks:4828): GLib-GObject-WARNING **: invalid (NULL) pointer instance

(gnome-disks:4828): GLib-GObject-CRITICAL **:
g_signal_handlers_disconnect_matched: assertion 'G_TYPE_CHECK_INSTANCE
(instance)' failed

(gnome-disks:4828): GLib-GObject-CRITICAL **: g_object_unref:
assertion 'G_IS_OBJECT (object)' failed

(gnome-disks:4828): GNOME-Disks-ERROR **: Error getting udisks client:
Could not connect: No such file or directory
Trace/breakpoint trap

Recompiling of gnome-disk-utility does not help anyway.
The package compiles almost nomarly but somewhere in the middle reports:

gducreatediskimagedialog.c:498:7: warning:
'gtk_dialog_get_action_area' is deprecated (declared at
/usr/include/gtk-3.0/gtk/gtkdialog.h:205) [-Wdeprecated-declarations]
   gtk_button_box_set_child_secondary (GTK_BUTTON_BOX
(gtk_dialog_get_action_area (GTK_DIALOG (dialog))),
   ^
gducreatediskimagedialog.c: In function 'check_overwrite':
gducreatediskimagedialog.c:933:3: warning:
'gtk_dialog_set_alternative_button_order' is deprecated (declared at
/usr/include/gtk-3.0/gtk/gtkdialog.h:187) [-Wdeprecated-declarations]
   gtk_dialog_set_alternative_button_order (GTK_DIALOG (dialog),
   ^
gdupasswordstrengthwidget.c: In function
'gdu_password_strength_widget_constructed':
gdupasswordstrengthwidget.c:205:7: warning: 'gtk_misc_set_alignment'
is deprecated (declared at
/usr/include/gtk-3.0/gtk/deprecated/gtkmisc.h:72)
[-Wdeprecated-declarations]
   gtk_misc_set_alignment (GTK_MISC (label), 0, 0.5);
   ^
gdupasswordstrengthwidget.c:205:7: warning: 'gtk_misc_get_type' is
deprecated (declared at
/usr/include/gtk-3.0/gtk/deprecated/gtkmisc.h:70)
[-Wdeprecated-declarations]

Moreover, I cannot print any more. (Have not checked it more than once, though.)

Any ideas, except that I should not update the system on December 13?

I use xfce4 but still on the default/linux/amd64/13.0/desktop/gnome profile.

The output of the system update on that day was as following:

# emerge --update --deep --with-bdeps=y --newuse --backtrack=90 --ask world

These are the packages that would be merged, in order:

Calculating dependencies... done!
[ebuild   R] dev-lang/python-exec-2.0.1-r1  PYTHON_TARGETS="(-jython2_5%*)"
[ebuild U  ] sys-apps/busybox-1.24.1 [1.23.1-r1]
[ebuild U  ] x11-proto/xproto-7.0.28 [7.0.27]
[ebuild U  ] x11-libs/libdrm-2.4.65 [2.4.59] USE="-valgrind%"
VIDEO_CARDS="-amdgpu%"
[ebuild U  ] x11-proto/kbproto-1.0.7 [1.0.6-r1]
[ebuild U  ] x11-proto/randrproto-1.5.0 [1.4.0-r1]
[ebuild U  ] media-fonts/font-util-1.3.1 [1.3.0]
[ebuild U  ] x11-libs/libpciaccess-0.13.4 [0.13.3]
[ebuild U  ] x11-libs/libXdmcp-1.1.2 [1.1.1-r1]
[ebuild U  ] x11-libs/libfontenc-1.1.3 [1.1.2]
[ebuild U  ] x11-base/xorg-drivers-1.17 [1.16]
VIDEO_CARDS="-amdgpu% (-tegra)"
[ebuild  N ] app-misc/c_rehash-1.7-r1
[ebuild U  ] dev-libs/openssl-1.0.2e [1.0.2d]
[blocks b  ] =app-eselect/eselect-opengl-1.3.0
(">=app-eselect/eselect-opengl-1.3.0" is blocking
x11-base/xorg-server-1.16.4)
[ebuild  rR] x11-drivers/xf86-video-ati-7.5.0
[ebuild U  ] x11-drivers/xf86-input-evdev-2.9.2 [2.9.1]
[ebuild U  ] media-libs/freeglut-3.0.0 [2.8.1-r1]
[ebuild U  ] dev-java/icedtea-bin-7.2.6.3 [7.2.6.2]
[ebuild U  ] sys-apps/openrc-0.18.4 [0.17]
[ebuild U  ] www-client/chromium-47.0.2526.80 [46.0.2490.86]
USE="hangouts%* (-gtk3) (-system-ffmpeg) -widevine%"

The following packages are causing rebuilds:

  (x11-libs/libxcb-1.11.1:0/1.11.1::gentoo, ebuild scheduled for
merge) causes rebuilds for:
(x11-libs/xcb-util-renderutil-0.3.9-r1:0/0::gentoo, ebuild
scheduled for merge)
(x11-libs/xcb-util-wm-0.4.1-r1:0/0::gentoo, ebuild scheduled for merge)
(x11-libs/xcb-util-image-0.4.0:0/0::gentoo, ebuild scheduled for merge)
(x11-libs/xcb-util-0.4.0:0/0::gentoo, ebuild scheduled for merge)
(x11-libs/xcb-util-keysyms-0.4.0:0/0::gentoo, ebuild scheduled for merge)
  

[gentoo-user] Re: Heads up for those who use grub2

[Grant Edwards]

On 2015-12-19, Mick  wrote:

> http://hmarco.org/bugs/CVE-2015-8370-Grub2-authentication-bypass.html

If somebody can touch your computer while it's booting, the game's
over anyway...

-- 
Grant

Re: [gentoo-user] Recommended pseudo-hardware for QEMU guest machine?

[wabenbau]

waltd...@waltdnes.org wrote:

>   I'm now at the configuring-the-kernel stage of the Gentoo guest
> install.  I had originally expected to pull in the .config from the
> host machine, make a few tweaks, and get going.  However, it appears
> that multiple video and sound and network cards are supported, none
> of which match those on the host.  Which ones do people recommend
> selecting?
> 

That's what I'm using to start qemu:

qemu-system-x86_64 -machine accel=kvm -cpu host -m 4096 -enable-kvm -name vm-01 
-net nic,model=virtio -net user,hostfwd=tcp::2022-:22 -localtime -hda 
/path/to/image.qcow2 -display gtk -vga vmware

Since upgrade to >=qemu-2.4 I cannot use sound card emulation any more
because it either doesn't work at all (e.g. sb16, gus) or it freezes
the guest (e.g. ac97, hda).

Before I did the upgrade yesterday I used -soundhw ac97 and it worked
very well. Till now I have not found a solution for this problem.

I'm using a gtk window and not spice or vnc because this gives me the
best performance. And I'm using vmware as VGA card emulation because 
it gives me the highest resolution (2368x1770).

Performance is not super snappy but it's good enough for regular use
like browsing the net and so on.

--
Regards
wabe

Re: [gentoo-user] Recommended pseudo-hardware for QEMU guest machine?

[waltdnes]

On Sun, Dec 20, 2015 at 04:47:35AM +0100, waben...@gmail.com wrote
> waltd...@waltdnes.org wrote:
> 
> >   I'm now at the configuring-the-kernel stage of the Gentoo guest
> > install.  I had originally expected to pull in the .config from the
> > host machine, make a few tweaks, and get going.  However, it appears
> > that multiple video and sound and network cards are supported, none
> > of which match those on the host.  Which ones do people recommend
> > selecting?
> > 
> 
> That's what I'm using to start qemu:
> 
> qemu-system-x86_64 -machine accel=kvm -cpu host -m 4096 -enable-kvm -name 
> vm-01 -net nic,model=virtio -net user,hostfwd=tcp::2022-:22 -localtime -hda 
> /path/to/image.qcow2 -display gtk -vga vmware
> 
> Since upgrade to >=qemu-2.4 I cannot use sound card emulation any more
> because it either doesn't work at all (e.g. sb16, gus) or it freezes
> the guest (e.g. ac97, hda).
> 
> Before I did the upgrade yesterday I used -soundhw ac97 and it worked
> very well. Till now I have not found a solution for this problem.
> 
> I'm using a gtk window and not spice or vnc because this gives me the
> best performance. And I'm using vmware as VGA card emulation because 
> it gives me the highest resolution (2368x1770).
> 
> Performance is not super snappy but it's good enough for regular use
> like browsing the net and so on.

  I intend to use it mostly for distcc.  I still have an ancient OS/2
machine with Galactic Civilizations on it.  Since I still have the CD
images,  I'll also try to get that going in QEMU one of these days.  My
LCD is 1920x1080, so performance is more important than humoungous
display size.

-- 
Walter Dnes 
I don't run "desktop environments"; I run useful applications

Re: [gentoo-user] Make QEMU guest visible to other machines on LAN

[wabenbau]

waltd...@waltdnes.org wrote:

>   Apologies if this is a duplicate/triplicate.  I don't think the
> first attempts got through.  Going through my backup provider
> (dialup) this time.
> 
>   I have QEMU installed on a 64-bit Gentoo machine.  I'm now
> installing a 32-bit Gentoo guest.  The "cdrom" (actually the minimal
> install ISO file) boots, and dhcpcd hands out IP address 10.0.2.15
> and gateway 10.0.2.2.  The install can connect to the outside world
> via the "links" browser and it can ssh into the host machine
> 192.168.123.249 and visa versa.  But the host ssh's into the guest
> install session via a port redirection into itself. ( ssh -p 
> localhost ) 
> 
>   For various reasons, I need another physical machine on my small
> home LAN to be able to talk directly to the 32-bit guest.  I've read
> the "Network setup" at http://www.linux-kvm.org/page/KvmOnGentoo
> Is the /etc/conf.d/net being reffered to, the one on the host or on
> the guest? The webpage doesn't say explicitly.

It seems like you are using the hostfwd option to redirect incoming 
connections to the host to the guest. Then it should also be able to 
connect from another physical machine to your guest.

Im using 

-net user,hostfwd=tcp::2022-:22

and can connect not only from the host but also from other physical
machines to the guest via ssh.

--
Regards
wabe

Re: [gentoo-user] [SOLVED] no network "eth0" after upgrade.

[J. Roeleveld]

On 20 December 2015 05:19:13 CET, Philip Webb  wrote:
>151219 the...@sys-concept.com wrote:
>> On 12/19/2015 05:56 PM, Neil Bothwick wrote:
>>> On Sat, 19 Dec 2015 19:02:54 -0500, Philip Webb wrote:
>Thelma> none   /proc/bus/usb   usbfs   defaults,devmode=0666  
>0 0
>> none   /proc   proc
>> defaults   0 0  
>Neil> You're trying to mount /proc/bus/usb before mounting /proc.
>> Systemd takes care of such things,  
>> but with Openrc local mounts are mounted in the order they appear in
>> fstab. Try switching the lines.  
>Philip> That looks like the explanation to me : has Thelma tried this ?
>Neil> Looking at it again, I don't think either of those entries
>> should be in fstab, certainly not the /proc one.
>Thelma> Good hint, thank you.  I've removed both lines from fstab:
>> and the system booted normally with openrc-0.18.4
>> Question, why isn't 'none /proc proc defaults 0 0' needed anymore ?
>> All my other systems have this line in fstab.
>
>My own Fstab has long had these lines :
>
>  # NB The next line is critical for boot!
>  none   /proc   proc   defaults   0 0
>
>When did this change & why ?  Does anyone know ?

It's been deprecated for a while.
Not sure when, think it was in a news item.

It certainly hasn't been in stage3 tarballs for the past 2 years. (Unless it 
reappeared suddenly)
I know this because that's when I installed my laptop and that line never was 
there.

--
Joost 
-- 
Sent from my Android device with K-9 Mail. Please excuse my brevity.

Re: [gentoo-user] [SOLVED] no network "eth0" after upgrade.

[Philip Webb]

151219 the...@sys-concept.com wrote:
> On 12/19/2015 05:56 PM, Neil Bothwick wrote:
>> On Sat, 19 Dec 2015 19:02:54 -0500, Philip Webb wrote:
Thelma> none   /proc/bus/usb   usbfs   defaults,devmode=0666   0 0
> none   /procproc
> defaults0 0  
Neil> You're trying to mount /proc/bus/usb before mounting /proc.
> Systemd takes care of such things,  
> but with Openrc local mounts are mounted in the order they appear in
> fstab. Try switching the lines.  
Philip> That looks like the explanation to me : has Thelma tried this ?
Neil> Looking at it again, I don't think either of those entries
> should be in fstab, certainly not the /proc one.
Thelma> Good hint, thank you.  I've removed both lines from fstab:
> and the system booted normally with openrc-0.18.4
> Question, why isn't 'none /proc proc defaults 0 0' needed anymore ?
> All my other systems have this line in fstab.

My own Fstab has long had these lines :

  # NB The next line is critical for boot!
  none   /proc   proc   defaults   0 0

When did this change & why ?  Does anyone know ?

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

Re: [gentoo-user] no network "eth0" after upgrade.

[Philip Webb]

151219 Neil Bothwick wrote:
> On Sat, 19 Dec 2015 13:59:38 -0700, the...@sys-concept.com wrote:
>> This is my fstab:
   ...
>> none   /proc/bus/usb   usbfs   defaults,devmode=0666   0 0
>> none   /proc   procdefaults0 0
> You're trying to mount /proc/bus/usb before mounting /proc.
> Systemd takes care of such things,
>> but with Openrc local mounts are mounted in the order they appear in fstab.
>> Try switching the lines.

That looks like the explanation to me : has Thelma tried this ?

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

Re: [gentoo-user] Hard drive noise

[wabenbau]

"J. Roeleveld"  wrote:

> >Since about three years I'm using four 3TB WD red HDs as storage
> >drives
> >
> >and I bought two more some months ago. No failures with all of these 
> >drives so far.
> >
> >--
> >Regards
> >wabe
> 
> I've got 16 3TB WD Reds running 24/7 for a little over 3 years. 
> Only had 1 failure (Smart complaining) in that time.
> 
> I find that decent odds.

That sounds good. It seems that it was a good decision to buy these
drives.

The time will show how reliable they are when the warranty period is 
up. :-)

--
Regards
wabe

Re: [gentoo-user] btrfs strategies (wasHard drive noise)

[J. Roeleveld]

On Sunday, December 20, 2015 02:15:33 AM cov...@ccs.covici.com wrote:
> cov...@ccs.covici.com wrote:
> > Rich Freeman  wrote:
> > > On Sat, Dec 19, 2015 at 2:56 AM,   wrote:
> > > > I was never able to get either zfs or btrfs to work correctly, zfs was
> > > > very vulnerable -- I forgot to export a zfs on a usb drive and got an
> > > > enless loop of processes untill I rebooted.  Btrfs never did work for
> > > > me, I created a pool,  copied my root file system, usr and var into
> > > > ssubvolumes, and copied my files, but when I would boot into it,
> > > > everything was messed up, processes thought files were missing, very
> > > > strange.  So, how did you set up either one of those -- I would love
> > > > to
> > > > use it because I have ssds and I don't want to rely on their firmware
> > > > either.
> > > 
> > > Well, I don't have much personal experience with zfs, but the ZFS on
> > > Linux lead is a Gentoo dev, so you're in good company there all the
> > > same.  I personally use btrfs.
> > > 
> > > The obvious caveat is that it is still relatively experimental, and
> > > raid5/6 is VERY experimental.  I plan to convert to raid5/6 at a
> > > future date but am staying away from it for now (and a selling point
> > > of btrfs is that reshaping in-place is easy).
> > > 
> > > I can't really vouch for what went wrong with your migration.  It
> > > could be anything from a failure to preserve all your file attributes
> > > to something with btrfs itself or your bootloader config/etc.  It
> > > isn't hard to do a new install in btrfs though, and you can always
> > > mess with it in a VM, or even mess with doing migrations in a VM.
> > > 
> > > My btrfs install notes are at:
> > > https://docs.google.com/document/d/1VJlJyYLTZScta9a81xgKOIBjYsG3_VfxxmUS
> > > xG23Uxg/edit?usp=sharing (I still plan to merge this stuff into the
> > > handbook.  Maybe a good holiday project...  Oh, and if it isn't already
> > > obvious anybody can add comments and half this list seems to have
> > > already done so.)
> > > 
> > > Oh, for a boot image I tend to use system rescue CD since it has all
> > > the necessary userspace and is gentoo-based (and you can always emerge
> > > --sync and install whatever you need inside it).  I tend to use the
> > > alternate kernel since it is newer, and with btrfs newer tends to be
> > > better.  In production I'm currently on 3.18 eyeing an upgrade to 4.1.
> > > I tend to stay on the latest longterm, but not when they are first
> > > declared as longterm.  That seems to be the sweet spot for getting
> > > btrfs features and bugfixes, but not getting as many of the
> > > regressions.  I use grub2/dracut to boot, and that is in my guide.
> > > 
> > > If you follow those notes for a stage3 install it should "just work."
> > > If you want to mess around I suggest just doing a vanilla install on a
> > > VM once to validate that it works for you and then tweak from a
> > > position of strength.
> > 
> > Thanks.  I will check out your notes and figure out something -- it was
> > definitely strange.  I have a vm I can play with -- its older, but I can
> > bring it up to date and see what happens.
> > 
> > Thanks again.
> 
> One thing I was thinking of -- since I like separate file systems for
> each major directory i.e. separate /usr, /var, /home, /tmp and even
> /var/tmp/portage, I thought I would make btrfs file systems using lvm.
> The advantage is that I use lvm already, so this would be easy for me to
> do and safer in case one of them goes south and easier to control  space
> allocation.  The only disadvantage I can see is if its a performance
> hit, does anyone have any knowledge of that is true?

I only played around with ZFS so far, but I believe the same holds true for 
BTRFS.

These new filesystems should really be handed control of the entire disk as 
they already include LVM-like functionality.
You can create subvolumes and limit those to different sizes if you so desire.

When using an additional layer between ZFS/BTRFS and the discs, you will loose 
performance with no gain in flexibility.

--
Joost

Re: [gentoo-user] no network "eth0" after upgrade.

[Neil Bothwick]

On Sat, 19 Dec 2015 19:02:54 -0500, Philip Webb wrote:

> >> none   /proc/bus/usb   usbfs   defaults,devmode=0666   0 0
> >> none   /proc proc
> >> defaults 0 0  
> > You're trying to mount /proc/bus/usb before mounting /proc.
> > Systemd takes care of such things,  
> >> but with Openrc local mounts are mounted in the order they appear in
> >> fstab. Try switching the lines.  
> 
> That looks like the explanation to me : has Thelma tried this ?

Looking at it again, I don't think either of those entries should be in
fstab, certainly not the /proc one.


-- 
Neil Bothwick

If Wile E. Coyote had enough money to buy all that ACME crap, why didn't
he just buy dinner?


pgp3p3bBi5io9.pgp
Description: OpenPGP digital signature

Re: [gentoo-user] Make QEMU guest visible to other machines on LAN

[waltdnes]

On Sat, Dec 19, 2015 at 08:31:35PM +0100, Simon Thelen wrote
> On 15-12-19 at 14:21, waltd...@waltdnes.org wrote:
> >   Apologies if this is a duplicate/triplicate.  I don't think the first
> > attempts got through.  Going through my backup provider (dialup) this
> > time.
> [..]
> >   For various reasons, I need another physical machine on my small home
> > LAN to be able to talk directly to the 32-bit guest.  I've read the
> > "Network setup" at http://www.linux-kvm.org/page/KvmOnGentooIs the
> > /etc/conf.d/net being reffered to, the one on the host or on the guest?
> > The webpage doesn't say explicitly.
> host
> > 
> >   I also don't understand how other machines will be able to
> > differentiate between the host and the guest.
> The idea is to create a bridge device, and then bridge the guests
> tun/tap device with the bridge and your ethernet device. This way they
> all appear transparently on the network.
> 
> The net config on that website should be fine for the host, and then
> pass these command options to qemu
> -device virtio-net-pci,netdev=hn0,mac="52:54:00:12:34:56" -netdev 
> bridge,id=hn0,br=brkvm
> 
> You can change the mac address to your liking.

  Thanks.  I'll try that as soon as I get the guest booting by itself.

-- 
Walter Dnes 
I don't run "desktop environments"; I run useful applications

Re: [gentoo-user] [SOLVED] no network "eth0" after upgrade.

[thelma]

On 12/19/2015 05:56 PM, Neil Bothwick wrote:
> On Sat, 19 Dec 2015 19:02:54 -0500, Philip Webb wrote:
> 
 none   /proc/bus/usb   usbfs   defaults,devmode=0666   0 0
 none   /proc proc
 defaults 0 0  
>>> You're trying to mount /proc/bus/usb before mounting /proc.
>>> Systemd takes care of such things,  
 but with Openrc local mounts are mounted in the order they appear in
 fstab. Try switching the lines.  
>>
>> That looks like the explanation to me : has Thelma tried this ?
> 
> Looking at it again, I don't think either of those entries should be in
> fstab, certainly not the /proc one.

Good hint, thank you.
I've removed both lines from fstab:
# Scanner
none   /proc/bus/usb   usbfs   defaults,devmode=0666   0 0
none /proc proc defaults 0 0

and the system booted normally with  openrc-0.18.4

Question, why isn't the line:
none /proc proc defaults 0 0
needed anymore?

All my other systems have this line in fstab.

--
Thelma

[gentoo-user] Recommended pseudo-hardware for QEMU guest machine?

[waltdnes]

  I'm now at the configuring-the-kernel stage of the Gentoo guest
install.  I had originally expected to pull in the .config from the host
machine, make a few tweaks, and get going.  However, it appears that
multiple video and sound and network cards are supported, none of which
match those on the host.  Which ones do people recommend selecting?

-- 
Walter Dnes 
I don't run "desktop environments"; I run useful applications

Re: [gentoo-user] [SOLVED] no network "eth0" after upgrade.

[thelma]

On 12/19/2015 09:19 PM, Philip Webb wrote:
> 151219 the...@sys-concept.com wrote:
>> On 12/19/2015 05:56 PM, Neil Bothwick wrote:
>>> On Sat, 19 Dec 2015 19:02:54 -0500, Philip Webb wrote:
> Thelma> none   /proc/bus/usb   usbfs   defaults,devmode=0666   0 0
>> none   /proc   proc
>> defaults   0 0  
> Neil> You're trying to mount /proc/bus/usb before mounting /proc.
>> Systemd takes care of such things,  
>> but with Openrc local mounts are mounted in the order they appear in
>> fstab. Try switching the lines.  
> Philip> That looks like the explanation to me : has Thelma tried this ?
> Neil> Looking at it again, I don't think either of those entries
>> should be in fstab, certainly not the /proc one.
> Thelma> Good hint, thank you.  I've removed both lines from fstab:
>> and the system booted normally with openrc-0.18.4
>> Question, why isn't 'none /proc proc defaults 0 0' needed anymore ?
>> All my other systems have this line in fstab.
> 
> My own Fstab has long had these lines :
> 
>   # NB The next line is critical for boot!
>   none   /proc   proc   defaults   0 0
> 
> When did this change & why ?  Does anyone know ?

I would like to know that too!

Indeed I had same comments on some of my other systems, but it seems to
me this line is no longer needed.  I'm compiling two other backup system
and commented these line out from fstab, let see what happens.

--
Thelma

Re: [gentoo-user] Hard drive noise

[Rich Freeman]

On Sat, Dec 19, 2015 at 5:12 AM, Thomas Mueller
 wrote:
>
> Now I am considering an external hard drive with eSATA, more suitable for OS 
> installation (Linux, NetBSD, FreeBSD, Haiku?) than USB 3.0.  Only brand I 
> find is Micronet Fantom (GForce), or use Seagate NAS hard drive in an 
> enclosure with eSATA.

I use a cheap external "enclosure" with a port replicator.  The
replicator part is sometimes problematic - sometimes one drive or the
other isn't recognized and I need to power-cycle (which means
unmounting both drives before touching either).  But, otherwise it
works fine, and lets me just use whatever internal drive I want.

I use it for a few purposes:
1.  Ability to plug in external drives for offline storage (vs burning
tons of DVDs).  I had a growing collection of smaller drives I'd
replaced anyway, and I use them in RAID1 pairs.  Reminds me that I
should scrub them soon...

2.  Ability to easily hot-swap for drive failures.  When I get a RAID
failure I can plug a new drive into the enclosure as soon as I have it
and rebuild the array, which gets me back into full redundancy sooner.
Then at a convenient point I'll swap the drive into the internal bay.

>
> I really can't see why USB 3.0 is so more widely available than eSATA when 
> eSATA seems superior as far as I can tell.
>

I suspect it is the ease-of-use factor.  USB external drives were more
common than eSATA back when USB meant USB 2.0 and eSATA was just as
good as it is today.  Clearly performance wasn't the deciding factor
here.

I will say that SATA port replicators seem finicky, at least under
Linux.  With USB it is all idiot-proof.  With SATA of any kind I end
up figuring out how many PCI cards I can jam into my PC with as many
ports each as possible if I want a large number of drives.  Backblaze
uses port replicators, but they've basically tailored their hardware
to a single purpose so they're using the motherboard+SATA+replicator
design that is optimal for their needs.

-- 
Rich

Re: [gentoo-user] Hard drive noise

[Rich Freeman]

On Sat, Dec 19, 2015 at 2:56 AM,   wrote:
>
> I was never able to get either zfs or btrfs to work correctly, zfs was
> very vulnerable -- I forgot to export a zfs on a usb drive and got an
> enless loop of processes untill I rebooted.  Btrfs never did work for
> me, I created a pool,  copied my root file system, usr and var into
> ssubvolumes, and copied my files, but when I would boot into it,
> everything was messed up, processes thought files were missing, very
> strange.  So, how did you set up either one of those -- I would love to
> use it because I have ssds and I don't want to rely on their firmware
> either.

Well, I don't have much personal experience with zfs, but the ZFS on
Linux lead is a Gentoo dev, so you're in good company there all the
same.  I personally use btrfs.

The obvious caveat is that it is still relatively experimental, and
raid5/6 is VERY experimental.  I plan to convert to raid5/6 at a
future date but am staying away from it for now (and a selling point
of btrfs is that reshaping in-place is easy).

I can't really vouch for what went wrong with your migration.  It
could be anything from a failure to preserve all your file attributes
to something with btrfs itself or your bootloader config/etc.  It
isn't hard to do a new install in btrfs though, and you can always
mess with it in a VM, or even mess with doing migrations in a VM.

My btrfs install notes are at:
https://docs.google.com/document/d/1VJlJyYLTZScta9a81xgKOIBjYsG3_VfxxmUSxG23Uxg/edit?usp=sharing
(I still plan to merge this stuff into the handbook.  Maybe a good
holiday project...  Oh, and if it isn't already obvious anybody can
add comments and half this list seems to have already done so.)

Oh, for a boot image I tend to use system rescue CD since it has all
the necessary userspace and is gentoo-based (and you can always emerge
--sync and install whatever you need inside it).  I tend to use the
alternate kernel since it is newer, and with btrfs newer tends to be
better.  In production I'm currently on 3.18 eyeing an upgrade to 4.1.
I tend to stay on the latest longterm, but not when they are first
declared as longterm.  That seems to be the sweet spot for getting
btrfs features and bugfixes, but not getting as many of the
regressions.  I use grub2/dracut to boot, and that is in my guide.

If you follow those notes for a stage3 install it should "just work."
If you want to mess around I suggest just doing a vanilla install on a
VM once to validate that it works for you and then tweak from a
position of strength.

-- 
Rich

Re: [gentoo-user] Hard drive noise

[J. Roeleveld]

On 19 December 2015 04:05:25 CET, waben...@gmail.com wrote:
>Dale  wrote:
>
>> I think I'm leery of all drives now.  I've had WDs fail, Seagate and
>> some other brand.  I just keep buying bad stuff.  :-(   I'm glad I
>> don't have to buy pacemakers.   :/ 
>
>It seems that many many years ago HDs were more reliable then today. 
>I have five 4GB IBM SCSI HDs and one 40GB Seagate IDE HD in my cupboard
>
>that I've used 24/7 for about 8 years. They were still intact when I 
>replaced  them.
>But its hardly surprising that a drive with a capacity of some TB has 
>a higher risk of failure than a drive with a capacity of 4GB or 40GB.
>
>Since about three years I'm using four 3TB WD red HDs as storage drives
>
>and I bought two more some months ago. No failures with all of these 
>drives so far.
>
>--
>Regards
>wabe

I've got 16 3TB WD Reds running 24/7 for a little over 3 years. 
Only had 1 failure (Smart complaining) in that time.

I find that decent odds.

--
Joost
-- 
Sent from my Android device with K-9 Mail. Please excuse my brevity.

Re: [gentoo-user] Hard drive noise

[Thomas Mueller]

> I've got 16 3TB WD Reds running 24/7 for a little over 3 years.
> Only had 1 failure (Smart complaining) in that time.

> I find that decent odds.

> Joost

I bought a WD Green 3 TB hard drive in May 2011, warranty was then 3 years.  It 
went bad with errors after 34 months.

I was able to get a warranty replacement after much hassle, with the warranty 
on such drives down to 2 years.

That replacement hard drive went bad in about seven months, strange sounds 
reminiscent of a dialup modem, drive was no longer recognized by the computer.

At nearly the same time as the WD Green failure, a 3 TB My Book Essential 3 TB 
USB 3.0 hard drive, ordered at the same time as the WD Green drive, went 
somewhat bad with errors, but the warranty on that was 2 years.

Needing a hard drive for another computer (May 2013), and not trusting WD or 
"Green", I ordered a Seagate NAS 4 TB hard drive, figuring increased 
reliability compared to Barracuda or Desktop was worth the modest additional 
cost.

That drive is still good as far as I can tell.

Now I am considering an external hard drive with eSATA, more suitable for OS 
installation (Linux, NetBSD, FreeBSD, Haiku?) than USB 3.0.  Only brand I find 
is Micronet Fantom (GForce), or use Seagate NAS hard drive in an enclosure with 
eSATA.

I really can't see why USB 3.0 is so more widely available than eSATA when 
eSATA seems superior as far as I can tell.

Tom

Re: perl ssl was:Re: [gentoo-user] dovecot imap-login

[jens wefer]

Am Mon, 14 Dec 2015 08:50:29 +0100
schrieb jens wefer :

> Am Sat, 12 Dec 2015 23:09:20 +0100
> schrieb jens wefer :
> 
> > Am Sat, 12 Dec 2015 17:53:04 +
> > schrieb Stroller :
> > 
> > > 
> > > > On Sat, 12 December 2015, at 3:08 a.m., jens wefer
> > > >  wrote:
> > > > 
> > > > I set up a mail server, postfix/dovecot, ssl required.
> > > > test with mail-client, all ok
> > > > when I try to copy mails with imapsync (gentoo) comes timeout,
> > > > and imapsync will login again.
> > > > with each new login, a new process imap-login is generated.
> > > 
> > > Sorry if this is a dumb question, but how do you know it's timing
> > > out?
> > > 
> > > Could it just be slow, as it has to compile loads of messages in
> > > its first run?
> > > 
> > > Looks like dovecot has a 30 minute timeout. [1]
> > > 
> > > An old message on the Dovecot mailing list [2] suggests to set
> > > "verbose_proctitle = yes" in config to see why each process is
> > > open.
> > > 
> > > It also suggests using high-performance mode, rather that the
> > > default. 
> > > 
> > > Stroller.
> > 
> > timeout comes from imapsync (default timeout 120 sec).
> > after 10 minutes then running 5 Dovecot processes which want 100%
> > CPU time. mail logfile:
> > imap-login: Login: user =  blablub, TLS ession, ..
> > 
> 
> I think that's a problem with perl.
> When I send an email with sendEmail comes SSLv3 Aler handshake
> failure. if I use a newer sendEmail version (1.56.5) comes
> Segnentation fault. when I start sendEmail on CentOS is everything ok.
> 


I send emails with email-client and sendEmail (win/centos).
mail.log
[...]: initializing the server-side TLS engine
[...]: connect from brumw.lxsbbshome.tld[192.168.0.15]
[...]: setting up TLS connection from brumw.lxsbbshome.tld[192.168.0.15]
[...]: brumw.lxsbbshome.tld[192.168.0.15]: TLS cipher list
"aNULL:-aNULL:ALL:!EXPORT:!LOW:+RC4:@STRENGTH" [...]:
SSL_accept:before/accept initialization [...]: SSL_accept:SSLv3 read
client hello A [...]: SSL_accept:SSLv3 write server hello A
[...]: SSL_accept:SSLv3 write certificate A
[...]: SSL_accept:SSLv3 write server done A
[...]: SSL_accept:SSLv3 flush data
[...]: SSL_accept:SSLv3 read client certificate A
[...]: SSL_accept:SSLv3 read client key exchange A
[...]: SSL_accept:SSLv3 read certificate verify A
[...]: SSL_accept:SSLv3 read finished A
[...]: brumw.lxsbbshome.tld[192.168.0.15]: Issuing session ticket, key
expiration: 1450478594 [...]: SSL_accept:SSLv3 write session ticket A
[...]: SSL_accept:SSLv3 write change cipher spec A
[...]: SSL_accept:SSLv3 write finished A
[...]: SSL_accept:SSLv3 flush data
[...]: Anonymous TLS connection established from
brumw.lxsbbshome.tld[192.168.0.15]: TLSv1.2 with cipher
AES128-GCM-SHA256 (128/128 bits) [...]: AFC46282149:
client=brumw.lxsbbshome.tld[192.168.0.15]

when I send email with sendEmail from gentoo-client it comes handshake
error mail.log
[...]: initializing the server-side TLS engine
[...]: connect from robin.lxsbbshome.tld[192.168.0.17]
[...]: setting up TLS connection from robin.lxsbbshome.tld[192.168.0.17]
[...]: robin.lxsbbshome.tld[192.168.0.17]: TLS cipher list
"aNULL:-aNULL:ALL:!EXPORT:!LOW:+RC4:@STRENGTH" [...]:
SSL_accept:before/accept initialization [...]: SSL3 alert
write:fatal:handshake failure [...]: SSL_accept:error in error
[...]: SSL_accept:error in error
[...]: SSL_accept error from robin.lxsbbshome.tld[192.168.0.17]: -1
[...]: warning: TLS library problem: error:1408A10B:SSL
routines:ssl3_get_client_hello:wrong version number:s3_srvr.c:960:
[...]: lost connection after STARTTLS from
robin.lxsbbshome.tld[192.168.0.17] [...]: disconnect from
robin.lxsbbshome.tld[192.168.0.17] ehlo=1 starttls=0/1 commands=1/2

sendEmail.log
[...]: DEBUG => Connecting to rosalie.lxsbbshome.tld:25
[...]: DEBUG => My IP address is: 192.168.0.17
[...]: DEBUG => evalSMTPresponse() - Checking for SMTP success or error
status in the message: 220 rosalie.lxsbbshome.tld ESMTP Postfix [...]:
DEBUG => evalSMTPresponse() - Found SMTP success code: 220 [...]:
SUCCESS => Received:220 rosalie.lxsbbshome.tld ESMTP Postfix
[...]: INFO => Sending: EHLO robin.lxsbbshome.tld [...]: DEBUG
=> evalSMTPresponse() - Checking for SMTP success or error status in
the message: 250-rosalie.lxsbbshome.tld, 250-PIPELINING, 250-SIZE
1024, 250-VRFY, 250-ETRN, 250-STARTTLS, 250-AUTH PLAIN,
250-ENHANCEDSTATUSCODES, 250-8BITMIME, 250-DSN, 250 SMTPUTF8 [...]:
DEBUG => evalSMTPresponse() - Found SMTP success code: 250 [...]:
SUCCESS => Received:250-rosalie.lxsbbshome.tld,
250-PIPELINING, 250-SIZE 1024, 250-VRFY, 250-ETRN, 250-STARTTLS,
250-AUTH PLAIN, 250-ENHANCEDSTATUSCODES, 250-8BITMIME, 250-DSN, 250
SMTPUTF8 [...]: DEBUG => The remote SMTP server supports TLS :) [...]:
DEBUG => Starting TLS [...]: INFO => Sending:   STARTTLS [...]:
DEBUG => evalSMTPresponse() - Checking for SMTP success or error status
in 

Re: [gentoo-user] no network "eth0" after upgrade.

[Neil Bothwick]

On Sat, 19 Dec 2015 13:59:38 -0700, the...@sys-concept.com wrote:

> This is my fstab:
> /dev/hda1 /boot   ext2
> noauto,noatime1
> 1 /dev/hda3   /   ext3
> noatime   0 1 /dev/hda2
> none  swapsw  0
> 0 /dev/hda4   /home   ext3
> noatime   0 1
> 
> /dev/hdd  /mnt/cdrom  auto
> noauto,ro,users   0
> 0 /dev/hdd/mnt/dvdr   auto
> noauto,users0 0
> 
> # Scanner
> none   /proc/bus/usb   usbfs   defaults,devmode=0666   0 0
> 
> none  /proc   proc
> defaults  0 0

Just a wild guess, but you're trying to mount /proc/bus/usb before
mounting /proc. systemd takes care of such things but I think with openrc
local mounts are mounted in the order they appear in fstab. Try switching
the lines.


-- 
Neil Bothwick

"I'm Not Sure If I'm Homosexual", Said Tom, Half In Earnest.


pgpO4GtYJ5_Fa.pgp
Description: OpenPGP digital signature

[gentoo-user] btrfs strategies (wasHard drive noise)

[covici]

cov...@ccs.covici.com wrote:

> Rich Freeman  wrote:
> 
> > On Sat, Dec 19, 2015 at 2:56 AM,   wrote:
> > >
> > > I was never able to get either zfs or btrfs to work correctly, zfs was
> > > very vulnerable -- I forgot to export a zfs on a usb drive and got an
> > > enless loop of processes untill I rebooted.  Btrfs never did work for
> > > me, I created a pool,  copied my root file system, usr and var into
> > > ssubvolumes, and copied my files, but when I would boot into it,
> > > everything was messed up, processes thought files were missing, very
> > > strange.  So, how did you set up either one of those -- I would love to
> > > use it because I have ssds and I don't want to rely on their firmware
> > > either.
> > 
> > Well, I don't have much personal experience with zfs, but the ZFS on
> > Linux lead is a Gentoo dev, so you're in good company there all the
> > same.  I personally use btrfs.
> > 
> > The obvious caveat is that it is still relatively experimental, and
> > raid5/6 is VERY experimental.  I plan to convert to raid5/6 at a
> > future date but am staying away from it for now (and a selling point
> > of btrfs is that reshaping in-place is easy).
> > 
> > I can't really vouch for what went wrong with your migration.  It
> > could be anything from a failure to preserve all your file attributes
> > to something with btrfs itself or your bootloader config/etc.  It
> > isn't hard to do a new install in btrfs though, and you can always
> > mess with it in a VM, or even mess with doing migrations in a VM.
> > 
> > My btrfs install notes are at:
> > https://docs.google.com/document/d/1VJlJyYLTZScta9a81xgKOIBjYsG3_VfxxmUSxG23Uxg/edit?usp=sharing
> > (I still plan to merge this stuff into the handbook.  Maybe a good
> > holiday project...  Oh, and if it isn't already obvious anybody can
> > add comments and half this list seems to have already done so.)
> > 
> > Oh, for a boot image I tend to use system rescue CD since it has all
> > the necessary userspace and is gentoo-based (and you can always emerge
> > --sync and install whatever you need inside it).  I tend to use the
> > alternate kernel since it is newer, and with btrfs newer tends to be
> > better.  In production I'm currently on 3.18 eyeing an upgrade to 4.1.
> > I tend to stay on the latest longterm, but not when they are first
> > declared as longterm.  That seems to be the sweet spot for getting
> > btrfs features and bugfixes, but not getting as many of the
> > regressions.  I use grub2/dracut to boot, and that is in my guide.
> > 
> > If you follow those notes for a stage3 install it should "just work."
> > If you want to mess around I suggest just doing a vanilla install on a
> > VM once to validate that it works for you and then tweak from a
> > position of strength.
> 
> Thanks.  I will check out your notes and figure out something -- it was
> definitely strange.  I have a vm I can play with -- its older, but I can
> bring it up to date and see what happens.
> 
> Thanks again.

One thing I was thinking of -- since I like separate file systems for
each major directory i.e. separate /usr, /var, /home, /tmp and even
/var/tmp/portage, I thought I would make btrfs file systems using lvm.
The advantage is that I use lvm already, so this would be easy for me to
do and safer in case one of them goes south and easier to control  space
allocation.  The only disadvantage I can see is if its a performance
hit, does anyone have any knowledge of that is true?

-- 
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] Re: systemd, libgudev and bug 552036

[Mike Gilbert]

On Fri, Dec 18, 2015 at 10:51 PM, Jonathan Callen  wrote:
> The python USE flag has been removed
> from newer stable versions of sys-apps/systemd (in favor of
> dev-python/python-systemd), but dev-python/python-systemd is not yet
> stable.

Thanks for catching that; I will file a stablereq right away.

Re: [gentoo-user] Hard drive noise

[J. Roeleveld]

On Saturday, December 19, 2015 10:12:22 AM Thomas Mueller wrote:
> > I've got 16 3TB WD Reds running 24/7 for a little over 3 years.
> > Only had 1 failure (Smart complaining) in that time.
> > 
> > I find that decent odds.
> > 
> > Joost
> 
> I bought a WD Green 3 TB hard drive in May 2011, warranty was then 3 years. 
> It went bad with errors after 34 months.

My experience with the WD Greens is similar. I don't trust them for important 
stuff.
They don't seem to manage 24/7 usage.

> I was able to get a warranty replacement after much hassle, with the
> warranty on such drives down to 2 years.
> 
> That replacement hard drive went bad in about seven months, strange sounds
> reminiscent of a dialup modem, drive was no longer recognized by the
> computer.
> 
> At nearly the same time as the WD Green failure, a 3 TB My Book Essential 3
> TB USB 3.0 hard drive, ordered at the same time as the WD Green drive, went
> somewhat bad with errors, but the warranty on that was 2 years.

I use the small 2.5" usb-drives from WD succesfully.
Only had 1 out of 8 die in the past 5 years. (Didn't bother with usb 
harddrives before then)

> Needing a hard drive for another computer (May 2013), and not trusting WD or
> "Green", I ordered a Seagate NAS 4 TB hard drive, figuring increased
> reliability compared to Barracuda or Desktop was worth the modest
> additional cost.
> 
> That drive is still good as far as I can tell.

Most manufacturers have good and bad drives.
Just make sure you pick the ones designed for your usage.
RAID setups don't like the aggressive powersaving implemented in "green" 
drives of any brand.

> Now I am considering an external hard drive with eSATA, more suitable for OS
> installation (Linux, NetBSD, FreeBSD, Haiku?) than USB 3.0.  Only brand I
> find is Micronet Fantom (GForce), or use Seagate NAS hard drive in an
> enclosure with eSATA.
> 
> I really can't see why USB 3.0 is so more widely available than eSATA when
> eSATA seems superior as far as I can tell.

I think USB 3.0 is cheaper and more common.
Only seen the occasional eSATA port on laptops and afaik, eSATA requires a 
seperate powersupply. USB can supply the power for the drive as well.

--
Joost

Re: [gentoo-user] Hard drive noise

[J. Roeleveld]

On Saturday, December 19, 2015 08:02:12 AM Rich Freeman wrote:
> On Sat, Dec 19, 2015 at 5:12 AM, Thomas Mueller
> 
>  wrote:
> > Now I am considering an external hard drive with eSATA, more suitable for
> > OS installation (Linux, NetBSD, FreeBSD, Haiku?) than USB 3.0.  Only
> > brand I find is Micronet Fantom (GForce), or use Seagate NAS hard drive
> > in an enclosure with eSATA.
> I use a cheap external "enclosure" with a port replicator.  The
> replicator part is sometimes problematic - sometimes one drive or the
> other isn't recognized and I need to power-cycle (which means
> unmounting both drives before touching either).  But, otherwise it
> works fine, and lets me just use whatever internal drive I want.

SATA and port replicators?
I've heard that for those to be reliable, you need a SAS controller.

> I use it for a few purposes:
> 1.  Ability to plug in external drives for offline storage (vs burning
> tons of DVDs).  I had a growing collection of smaller drives I'd
> replaced anyway, and I use them in RAID1 pairs.  Reminds me that I
> should scrub them soon...

I currently use 2.5" drives in hot-swap bays myself. External enclosures means 
similar amount of work swapping them, but with the added complexity and wiring 
when using external enclosures.

> 2.  Ability to easily hot-swap for drive failures.  When I get a RAID
> failure I can plug a new drive into the enclosure as soon as I have it
> and rebuild the array, which gets me back into full redundancy sooner.
> Then at a convenient point I'll swap the drive into the internal bay.
> 
> > I really can't see why USB 3.0 is so more widely available than eSATA when
> > eSATA seems superior as far as I can tell.
> I suspect it is the ease-of-use factor.  USB external drives were more
> common than eSATA back when USB meant USB 2.0 and eSATA was just as
> good as it is today.  Clearly performance wasn't the deciding factor
> here.

Power from the bus? (Eg. reducing the amount of cables)

> I will say that SATA port replicators seem finicky, at least under
> Linux.  With USB it is all idiot-proof.  With SATA of any kind I end
> up figuring out how many PCI cards I can jam into my PC with as many
> ports each as possible if I want a large number of drives.  Backblaze
> uses port replicators, but they've basically tailored their hardware
> to a single purpose so they're using the motherboard+SATA+replicator
> design that is optimal for their needs.

Backblaze actually wrote about which chipsets work together.
If you stick with those, it should work.

--
Joost

Re: [gentoo-user] Hard drive noise

[covici]

Rich Freeman  wrote:

> On Sat, Dec 19, 2015 at 2:56 AM,   wrote:
> >
> > I was never able to get either zfs or btrfs to work correctly, zfs was
> > very vulnerable -- I forgot to export a zfs on a usb drive and got an
> > enless loop of processes untill I rebooted.  Btrfs never did work for
> > me, I created a pool,  copied my root file system, usr and var into
> > ssubvolumes, and copied my files, but when I would boot into it,
> > everything was messed up, processes thought files were missing, very
> > strange.  So, how did you set up either one of those -- I would love to
> > use it because I have ssds and I don't want to rely on their firmware
> > either.
> 
> Well, I don't have much personal experience with zfs, but the ZFS on
> Linux lead is a Gentoo dev, so you're in good company there all the
> same.  I personally use btrfs.
> 
> The obvious caveat is that it is still relatively experimental, and
> raid5/6 is VERY experimental.  I plan to convert to raid5/6 at a
> future date but am staying away from it for now (and a selling point
> of btrfs is that reshaping in-place is easy).
> 
> I can't really vouch for what went wrong with your migration.  It
> could be anything from a failure to preserve all your file attributes
> to something with btrfs itself or your bootloader config/etc.  It
> isn't hard to do a new install in btrfs though, and you can always
> mess with it in a VM, or even mess with doing migrations in a VM.
> 
> My btrfs install notes are at:
> https://docs.google.com/document/d/1VJlJyYLTZScta9a81xgKOIBjYsG3_VfxxmUSxG23Uxg/edit?usp=sharing
> (I still plan to merge this stuff into the handbook.  Maybe a good
> holiday project...  Oh, and if it isn't already obvious anybody can
> add comments and half this list seems to have already done so.)
> 
> Oh, for a boot image I tend to use system rescue CD since it has all
> the necessary userspace and is gentoo-based (and you can always emerge
> --sync and install whatever you need inside it).  I tend to use the
> alternate kernel since it is newer, and with btrfs newer tends to be
> better.  In production I'm currently on 3.18 eyeing an upgrade to 4.1.
> I tend to stay on the latest longterm, but not when they are first
> declared as longterm.  That seems to be the sweet spot for getting
> btrfs features and bugfixes, but not getting as many of the
> regressions.  I use grub2/dracut to boot, and that is in my guide.
> 
> If you follow those notes for a stage3 install it should "just work."
> If you want to mess around I suggest just doing a vanilla install on a
> VM once to validate that it works for you and then tweak from a
> position of strength.

Thanks.  I will check out your notes and figure out something -- it was
definitely strange.  I have a vm I can play with -- its older, but I can
bring it up to date and see what happens.

Thanks again.

-- 
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