Re: recommended poudriere jail versions?

[Guido Falsi]

On 10/01/15 14:05, Mark Martinec wrote:
> 2015-10-01 10:32, Marko Cupać wrote:
>> what is the recommended poudriere jail version for building ports? So
>> far I was trying to be on latest binary patchlevel for every minor
>> version for both base system, poudriere jails and clients, but I ended
>> up with three jails just for amd64 (9.3, 10.1 and 10.2), where I need to
>> rebuild all the ports every time I patch poudriere jails. This is
>> starting to take too much of my time.
>>
>> I see that pkg.freebsd.org hosts just one set of ports per
>> architecture of major version. What is the OS version they are built
>> on? Are there any downsides in building all the ports for
>> 10.2- on 10.1-?
> 
> I used to have poudriere jails based on a minor version like you have,
> but ended up in a simplified setup, building ports only on 10.0-RELEASE
> and installing them on 10.1 or 10.2 and 10-STABLE. I think the
> official packages are also built based on 10.0-RELEASE .

Official packages are being build on 10.1-RELEASE at present. It's the
lowest supported minor in the 10 branch.

-- 
Guido Falsi 
___
freebsd-stable@freebsd.org mailing list
https://lists.freebsd.org/mailman/listinfo/freebsd-stable
To unsubscribe, send any mail to "freebsd-stable-unsubscr...@freebsd.org"

Re: recommended poudriere jail versions?

[Marko Cupać]

On Thu, 01 Oct 2015 14:05:32 +0200
Mark Martinec  wrote:

> 2015-10-01 10:32, Marko Cupać wrote:
> > what is the recommended poudriere jail version for building ports?
> > So far I was trying to be on latest binary patchlevel for every
> > minor version for both base system, poudriere jails and clients,
> > but I ended up with three jails just for amd64 (9.3, 10.1 and
> > 10.2), where I need to
> > rebuild all the ports every time I patch poudriere jails. This is
> > starting to take too much of my time.
> > 
> > I see that pkg.freebsd.org hosts just one set of ports per
> > architecture of major version. What is the OS version they are built
> > on? Are there any downsides in building all the ports for
> > 10.2- on 10.1-?
> 
> I used to have poudriere jails based on a minor version like you have,
> but ended up in a simplified setup, building ports only on
> 10.0-RELEASE and installing them on 10.1 or 10.2 and 10-STABLE. I
> think the official packages are also built based on 10.0-RELEASE .
> 
> This mostly works, except for a port like virtualbox-ose-kmod,
> which causes a kernel crash when built on 10.0-RELEASE and run
> on 10.2. So after each ports upgrade when noticing that pkg
> is reinstalling virtualbox-ose-kmod, I re-build this one from
> ports on a target host, otherwise the next reboot will end up
> crashing on loading a vboxdrv kernel module during startup.

Thanx for info. I'd like to give building everything on
10.1-RELEASE- a try, but I don't see the ability of poudriere
to specify patclevel when using ftp method - it always fetches the
latest one. Is this by design, to discourage building on potentially
vulnerable versions, or just non-implemented functionality?

Regards,
-- 
Before enlightenment - chop wood, draw water.
After  enlightenment - chop wood, draw water.

Marko Cupać
https://www.mimar.rs/
___
freebsd-stable@freebsd.org mailing list
https://lists.freebsd.org/mailman/listinfo/freebsd-stable
To unsubscribe, send any mail to "freebsd-stable-unsubscr...@freebsd.org"

em0 and em1 watchdog timed out

[Frank de Bot (lists)]

On a server I have 2 interfaces configured in a lagg. It seemed to be
running stable until now (about a week), both interfaces repeatable go
down after a watchdog timeout.

Output from /var/log/messages

Oct  2 10:15:11 nas kernel: em0: Watchdog timeout Queue[0]-- resetting
Oct  2 10:15:11 nas kernel: Interface is RUNNING and ACTIVE
Oct  2 10:15:11 nas kernel: em0: TX Queue 0 --
Oct  2 10:15:11 nas kernel: em0: hw tdh = 804, hw tdt = 905
Oct  2 10:15:11 nas kernel: em0: Tx Queue Status = -2147483648
Oct  2 10:15:11 nas kernel: em0: TX descriptors avail = 915
Oct  2 10:15:11 nas kernel: em0: Tx Descriptors avail failure = 0
Oct  2 10:15:11 nas kernel: em0: RX Queue 0 --
Oct  2 10:15:11 nas kernel: em0: hw rdh = 912, hw rdt = 911
Oct  2 10:15:11 nas kernel: em0: RX discarded packets = 0
Oct  2 10:15:11 nas kernel: em0: RX Next to Check = 912
Oct  2 10:15:11 nas kernel: em0: RX Next to Refresh = 911
Oct  2 10:15:11 nas kernel: em0: link state changed to DOWN
Oct  2 10:15:13 nas kernel: em0: link state changed to UP
Oct  2 10:15:13 nas devd: Executing '/etc/rc.d/dhclient quietstart em0'
Oct  2 10:21:56 nas kernel: em1: Interface stopped DISTRIBUTING,
possible flapping
Oct  2 10:21:59 nas kernel: em1: Watchdog timeout Queue[0]-- resetting
Oct  2 10:21:59 nas kernel: Interface is RUNNING and ACTIVE
Oct  2 10:21:59 nas kernel: em1: TX Queue 0 --
Oct  2 10:21:59 nas kernel: em1: hw tdh = 686, hw tdt = 806
Oct  2 10:21:59 nas kernel: em1: Tx Queue Status = -2147483648
Oct  2 10:21:59 nas kernel: em1: TX descriptors avail = 896
Oct  2 10:21:59 nas kernel: em1: Tx Descriptors avail failure = 0
Oct  2 10:21:59 nas kernel: em1: RX Queue 0 --
Oct  2 10:21:59 nas kernel: em1: hw rdh = 167, hw rdt = 166
Oct  2 10:21:59 nas kernel: em1: RX discarded packets = 0
Oct  2 10:21:59 nas kernel: em1: RX Next to Check = 167
Oct  2 10:21:59 nas kernel: em1: RX Next to Refresh = 166
Oct  2 10:21:59 nas kernel: em1: link state changed to DOWN
Oct  2 10:22:02 nas kernel: em1: link state changed to UP
Oct  2 10:22:02 nas devd: Executing '/etc/rc.d/dhclient quietstart em1'

it seems to occur when the interface are loaded. On the same server I
also have a PCI card, the same issue occurs here. em0 and em1 are both
onboard 80003ES2LAN. the PCI card is 82571EB.

I've found different threads, but I couldn't manage to find  a fix for this.

What else can be wrong?

Regards,

Frank de Bot
___
freebsd-stable@freebsd.org mailing list
https://lists.freebsd.org/mailman/listinfo/freebsd-stable
To unsubscribe, send any mail to "freebsd-stable-unsubscr...@freebsd.org"

Re: em0 and em1 watchdog timed out

[Jack Vogel]

You missed the all-important details: OS version, driver version.
And another question I can think of, do these interfaces watchdog if they
are not configured with lagg?

Cheers,

Jack


On Fri, Oct 2, 2015 at 10:37 AM, Frank de Bot (lists) 
wrote:

> On a server I have 2 interfaces configured in a lagg. It seemed to be
> running stable until now (about a week), both interfaces repeatable go
> down after a watchdog timeout.
>
> Output from /var/log/messages
>
> Oct  2 10:15:11 nas kernel: em0: Watchdog timeout Queue[0]-- resetting
> Oct  2 10:15:11 nas kernel: Interface is RUNNING and ACTIVE
> Oct  2 10:15:11 nas kernel: em0: TX Queue 0 --
> Oct  2 10:15:11 nas kernel: em0: hw tdh = 804, hw tdt = 905
> Oct  2 10:15:11 nas kernel: em0: Tx Queue Status = -2147483648
> Oct  2 10:15:11 nas kernel: em0: TX descriptors avail = 915
> Oct  2 10:15:11 nas kernel: em0: Tx Descriptors avail failure = 0
> Oct  2 10:15:11 nas kernel: em0: RX Queue 0 --
> Oct  2 10:15:11 nas kernel: em0: hw rdh = 912, hw rdt = 911
> Oct  2 10:15:11 nas kernel: em0: RX discarded packets = 0
> Oct  2 10:15:11 nas kernel: em0: RX Next to Check = 912
> Oct  2 10:15:11 nas kernel: em0: RX Next to Refresh = 911
> Oct  2 10:15:11 nas kernel: em0: link state changed to DOWN
> Oct  2 10:15:13 nas kernel: em0: link state changed to UP
> Oct  2 10:15:13 nas devd: Executing '/etc/rc.d/dhclient quietstart em0'
> Oct  2 10:21:56 nas kernel: em1: Interface stopped DISTRIBUTING,
> possible flapping
> Oct  2 10:21:59 nas kernel: em1: Watchdog timeout Queue[0]-- resetting
> Oct  2 10:21:59 nas kernel: Interface is RUNNING and ACTIVE
> Oct  2 10:21:59 nas kernel: em1: TX Queue 0 --
> Oct  2 10:21:59 nas kernel: em1: hw tdh = 686, hw tdt = 806
> Oct  2 10:21:59 nas kernel: em1: Tx Queue Status = -2147483648
> Oct  2 10:21:59 nas kernel: em1: TX descriptors avail = 896
> Oct  2 10:21:59 nas kernel: em1: Tx Descriptors avail failure = 0
> Oct  2 10:21:59 nas kernel: em1: RX Queue 0 --
> Oct  2 10:21:59 nas kernel: em1: hw rdh = 167, hw rdt = 166
> Oct  2 10:21:59 nas kernel: em1: RX discarded packets = 0
> Oct  2 10:21:59 nas kernel: em1: RX Next to Check = 167
> Oct  2 10:21:59 nas kernel: em1: RX Next to Refresh = 166
> Oct  2 10:21:59 nas kernel: em1: link state changed to DOWN
> Oct  2 10:22:02 nas kernel: em1: link state changed to UP
> Oct  2 10:22:02 nas devd: Executing '/etc/rc.d/dhclient quietstart em1'
>
> it seems to occur when the interface are loaded. On the same server I
> also have a PCI card, the same issue occurs here. em0 and em1 are both
> onboard 80003ES2LAN. the PCI card is 82571EB.
>
> I've found different threads, but I couldn't manage to find  a fix for
> this.
>
> What else can be wrong?
>
> Regards,
>
> Frank de Bot
> ___
> freebsd-stable@freebsd.org mailing list
> https://lists.freebsd.org/mailman/listinfo/freebsd-stable
> To unsubscribe, send any mail to "freebsd-stable-unsubscr...@freebsd.org"
>
___
freebsd-stable@freebsd.org mailing list
https://lists.freebsd.org/mailman/listinfo/freebsd-stable
To unsubscribe, send any mail to "freebsd-stable-unsubscr...@freebsd.org"

Re: recommended poudriere jail versions?

[Slawa Olhovchenkov]

On Thu, Oct 01, 2015 at 10:32:06AM +0200, Marko Cupa'c wrote:

> Hi,
> 
> what is the recommended poudriere jail version for building ports? So
> far I was trying to be on latest binary patchlevel for every minor
> version for both base system, poudriere jails and clients, but I ended
> up with three jails just for amd64 (9.3, 10.1 and 10.2), where I need to
> rebuild all the ports every time I patch poudriere jails. This is
> starting to take too much of my time.
> 
> I see that pkg.freebsd.org hosts just one set of ports per
> architecture of major version. What is the OS version they are built
> on? Are there any downsides in building all the ports for
> 10.2- on 10.1-?

Some packages (freeradius, for example) strict depends from openssl
version. Different openssl version in base (in jail and in target)
stop to run this application.
opensssl version in base can be changed by patchlevel too.
___
freebsd-stable@freebsd.org mailing list
https://lists.freebsd.org/mailman/listinfo/freebsd-stable
To unsubscribe, send any mail to "freebsd-stable-unsubscr...@freebsd.org"