8.0.3 RELEASE coming soon

[Raymond Matos]

Wait. I think I just overlooked a key piece of information: is your dynamic DNS provider zonedit and it's trying to access dyndns.com instead?

Yep thats the issues. I'm using zoneedit and it trying to access dyndns.com.

 

[gcooper]

Yep thats the issues. I'm using zoneedit and it trying to access dyndns.com.

DERP. I'll see if I can reproduce the issue. Worst comes to worst I'll supply a patch to you and consider doing a patch release in another couple weeks. I wish I had gotten this feedback sooner >_>.

 

[krakoliezene]

Dear gcooper,

What does it mean that /var/db/pkg was "nuked"? I have installed several jails for applications that are not present in the appliance. Will those still function? Otherwise I think we are sacrificing a lot of flexibility for some 7MB of RAM space. I understand that Freenas is an appliance, but that does not imply that anything extra is forbidden or am I missing something?

 

[gcooper]

Dear gcooper,

What does it mean that /var/db/pkg was "nuked"? I have installed several jails for applications that are not present in the appliance. Will those still function? Otherwise I think we are sacrificing a lot of flexibility for some 7MB of RAM space. I understand that Freenas is an appliance, but that does not imply that anything extra is forbidden or am I missing something?

/var/db/pkg is the path to the FreeBSD packaging metadata. It has zero value on FreeNAS with the way things are currently designed; if you're doing unsupported operations like pkg_add -r'ing stuff, then it's going to add more packages now; I highly advise not using FreeBSD 8.2 packages that weren't generated again the nanobsd build environment available in FreeNAS.

 

[krakoliezene]

/var/db/pkg is the path to the FreeBSD packaging metadata. It has zero value on FreeNAS with the way things are currently designed; if you're doing unsupported operations like pkg_add -r'ing stuff, then it's going to add more packages now; I highly advise not using FreeBSD 8.2 packages that weren't generated again the nanobsd build environment available in FreeNAS.

cooper,

Yes, I did use pkg_add, but just inside jails and stored on the ZFS pool, like the normal procedure seems to be when adding functionality to nanobsd. Now, I did all that, but only once and I don't have enough understanding of the whole FreeBSD concept to know for sure if upgrading to 8.0.3 RELEASE will get me in trouble. Could you give me any guidance?

 

[gcooper]

cooper,

Yes, I did use pkg_add, but just inside jails and stored on the ZFS pool, like the normal procedure seems to be when adding functionality to nanobsd. Now, I did all that, but only once and I don't have enough understanding of the whole FreeBSD concept to know for sure if upgrading to 8.0.3 RELEASE will get me in trouble. Could you give me any guidance?

No. /var/db/pkg was removed from the base system, not your jails. All space changes made only affect the FreeNAS base image.

 

[gcooper]

Ok, blah. Rebuilding on streetfighter.ixsystems.com... the issue with pathing and stuff with xvid wasn't being resolved consistently according to my observations, even after an xvid upgrade -_-. It seems to work interactively, but something doesn't work when I execute it via the 'deterministic build' script.

New ETA: sometime earlier on this afternoon (build.ix is faster iron than streetfighter.ix).

 

[Daisuke]

No. /var/db/pkg was removed from the base system, not your jails. All space changes made only affect the FreeNAS base image.

I'm confused. Are you suggesting we should install the new image or upgrading through GUI will have the same effect? I just upgraded through GUI without issues, but I want to make sure I have the proper image installed with the right directories.

Edit: As a side note, NFSv4 does not work properly (if the release note means version 4 is supported on FreeNAS). I tried to connect from a CentOS box and I get a version mismatch error. Is there anything that needs to be enabled into FreeNAS config files?

Code:

# rpcinfo -p pluto | grep nfs
    100003    2   udp   2049  nfs
    100003    3   udp   2049  nfs
    100003    2   tcp   2049  nfs
    100003    3   tcp   2049  nfs

Nothing about version 4.

 

[krakoliezene]

No. /var/db/pkg was removed from the base system, not your jails. All space changes made only affect the FreeNAS base image.

Yeah, I upgraded to 8.0.3. RELEASE and everything works like before!!

Thanks gcooper, for your help!

 

[joeschmuck]

Ok, blah. Rebuilding on streetfighter.ixsystems.com... the issue with pathing and stuff with xvid wasn't being resolved consistently according to my observations, even after an xvid upgrade -_-. It seems to work interactively, but something doesn't work when I execute it via the 'deterministic build' script.

New ETA: sometime earlier on this afternoon (build.ix is faster iron than streetfighter.ix).

Hope you figure it out.

 

[gcooper]

Hope you figure it out.

I will eventually. Basically it's a port bug somewhere with xvid where it fails on build.ixsystems.com because --prefix=/build instead of /usr/local >_>...

Code:

install: mkdir =build/dct/x86_asm: File exists
gmake: *** [dct/x86_asm/fdct_mmx_skal.o] Error 71
gmake: *** Waiting for unfinished jobs....
*** Error code 1

 

[joeschmuck]

I will eventually. Basically it's a port bug somewhere with xvid where it fails on build.ixsystems.com because --prefix=/build instead of /usr/local >_>...

Code:

install: mkdir =build/dct/x86_asm: File exists
gmake: *** [dct/x86_asm/fdct_mmx_skal.o] Error 71
gmake: *** Waiting for unfinished jobs....
*** Error code 1

I'm having no problems building. I'm running your mm-release from yesterday just fine on my real NAS. I haven't had time to build anything today. Unfortunately I'm not sure what the prefix thing means but if I need to dive into it I will learn quickly.

EDIT: I think I understand, it's the location of the files you are compiling? I am at /usr/local/freenas/, is that it?

 

[AndroidCat]

I tested the UPS correction (UPS low battery) and this worked for me. Thanks!

I can see however some flaw in shutdown logic..

'Shutdown timer' parameter should only apply in "UPS on battery" mode.

Currently in 8.0.3 UPS reaches "low battery" state (in my APC it is set to 10% of remaining charge) and then FN waits additional 'Shutdown timer' to close down the system.
In my opinion for "low battery" the timer should not matter and so the shutdown should happen immediatelly.
The timer should only apply to "UPS on battery" as in that case FN should wait a few minutes before going down.

1. Could you please check the shutdown script if you agree with my understanding of how the UPS should behave?
2. Is it possible to modify GUI to disable the timer box when "low on battery" is choosen (or print some info, warning, etc).

Of course I can adjust the timer to a very small value for "low on battery" case and forget it...

Once again, thanks for your work and Happy New Year!

upsc ups@localhost
battery.charge: 6 <- charge already gone a lot lower than expected due to timer set to 900s
battery.charge.low: 10 <- threshold when UPS generates "low on battery" warning


Jan 4 16:17:04 freenas upsmon[6008]: UPS ups on battery

Broadcast Message from root@freenas.local
(no tty) at 16:17 CST...

UPS ups on battery

Jan 4 16:53:19 freenas upsmon[6008]: UPS ups battery is low
Jan 4 16:53:19 freenas upsmon[6008]: Executing automatic power-fail shutdown

Broadcast Message from root@freenas.local
(no tty) at 16:53 CST...

UPS ups battery is low


Broadcast Message from root@freenas.local
(no tty) at 16:53 CST...

Executing automatic power-fail shutdown


Jan 4 16:53:19 freenas upsmon[6008]: Auto logout and shutdown proceeding

<waits 900s=15min for "real shutdown" LOL

*** FINAL System shutdown message from root@freenas.local ***
System going down IMMEDIATELY


Jan 4 17:08:19 freenas shutdown: power-down by root:
Jan 4 17:08:20 freenas upssched-cmd: issuing shutdown

 

[Daisuke]

Anyone has any issues with NFSv4? See my post.
I'm really looking forward on this, since NFS4 is available since FreeBSD 8.0. The performance enhancements and security are worth the upgrade, IMO.

 

[ProtoSD]

I have to say thanks on behalf of everyone that has been looking forward to the NTFS import being fixed! I know the NTFS implementation isn't perfect, but for people that want to import their data or copy something to/from a USB stick it's really nice to have it working!

THANKS!!!

 

[AndroidCat]

Anyone has any issues with NFSv4? See my post.

Release notes just mention:

Rsync - NFSv4 ACL support (see: r8375, r8414).

which means rsync protocol now understands NFSv4 access control list.

 

[gcooper]

Release notes just mention:

Rsync - NFSv4 ACL support (see: r8375, r8414).

which means rsync protocol now understands NFSv4 access control list.

Which means that rsync now groks Windows ACLs with the -A flag.

 

[gcooper]

Anyone has any issues with NFSv4? See my post.
I'm really looking forward on this, since NFS4 is available since FreeBSD 8.0. The performance enhancements and security are worth the upgrade, IMO.

NFSv4 was marked experimental until 9.x. Please be patient... it will come soon enough :).

 

[gcooper]

I'm having no problems building. I'm running your mm-release from yesterday just fine on my real NAS. I haven't had time to build anything today. Unfortunately I'm not sure what the prefix thing means but if I need to dive into it I will learn quickly.

EDIT: I think I understand, it's the location of the files you are compiling? I am at /usr/local/freenas/, is that it?

I'll look at it tonight. I am sure I'll have enough spare cycles to root out the issue this week and resolve it (build stuff is generally easy to fix).

 

[Raymond Matos]

Yep thats the issues. I'm using zoneedit and it trying to access dyndns.com.

Hey gcooper - i think the issue is with zoneedit.com, it looks like if the zone is up to date, it returns like its not a valid user/pass combo. So i edited the zone via the zoneedit.com UI, then started the service and it starts and updates the IP. Not sure why the system thinks the user/pass is invalid when the ip is up to date.