Bug Report Thread for B3 "hugo" RC Repository

[carl]

[Bug 5]
Samba daemon will not start if the wifi is enabled. However, starting samba first then enabling the wifi works fine.

The only thing I can see in the syslog is this:

Jul 9 13:16:53 serverb3 smbd[26120]: [2011/07/09 13:16:53.474838, 0] param/loadparm.c:7588(lp_do_parameter)
Jul 9 13:16:53 serverb3 smbd[26120]: Ignoring unknown parameter "interface"
Jul 9 13:16:53 serverb3 smbd[26120]: [2011/07/09 13:16:53.490481, 0] param/loadparm.c:7588(lp_do_parameter)
Jul 9 13:16:53 serverb3 smbd[26120]: Ignoring unknown parameter "interface"

This is from samba/log.all:

[2011/07/09 13:16:53.490278, 1] param/loadparm.c:6494(map_parameter)
Unknown parameter encountered: "interface"
[2011/07/09 13:16:53.490481, 0] param/loadparm.c:7588(lp_do_parameter)
Ignoring unknown parameter "interface"
[2011/07/09 13:16:53.622680, 0] lib/interface.c:542(load_interfaces)
WARNING: no network interfaces found
[2011/07/09 13:16:53.757681, 0] smbd/server.c:659(open_sockets_smbd)
open_sockets_smbd: No sockets available to bind to.
[2011/07/09 13:16:53.758655, 0] smbd/server.c:894(exit_server_common)
===============================================================
[2011/07/09 13:16:53.758832, 0] smbd/server.c:896(exit_server_common)
Abnormal server exit: open_sockets_smbd() failed
[2011/07/09 13:16:53.758936, 0] smbd/server.c:897(exit_server_common)
===============================================================
[2011/07/09 13:16:53.759666, 0] lib/util.c:1569(log_stack_trace)
BACKTRACE: 0 stack frames:
[2011/07/09 13:16:53.759871, 0] lib/fault.c:326(dump_core)
dumping core in /var/log/samba/cores/smbd

[EDIT]
OK, I've worked out what's happening here. When the wifi is enabled or disabled from the web interface the script is editing smb.conf but the script must be buggy (buggie?). I guess it's supposed to change the parameter "interfaces = eth1" to "interfaces = br0". However, it's adding "interface = br0" - note the missing "s" which is what is being complained about in the samba log. Also, it's leaving any previous "interfaces" parameter as well so I've got both these lines in smb.conf:

Code: Select all

interfaces = eth1
interface = br0

Even with the typo fixed I'm not sure you can have two lines the same. The correct formatting would be:

Code: Select all

interfaces = eth1 br0

However, just

Code: Select all

interfaces = br0

works fine.

Have been digging around in the code, and sadly, I've been unable to indicate anything that would set to "interface" instead of "interfaces"; I'll dig further, but can't guarantee I will find it in our code.

[Nrde]

logged in as admin is regarded in the album as being anonymous (was decided that the bubba admin should not touch the album)

But it is also true for user that is _also_ an admin user.

I have given myself admin rights, but that prevents me from using the album.

In my opinion "Admin" user should be prevented tampering the albums, but not normal users who happen to be also admins? Or is there something wrong with my anstall as I can't seem to be able to create albums, or at least see any if I'm logged in or out.

PS. Sorry to hijack this thread, please move my message as you see fit...

[RandomUsername]

[Bug 4]
When rebooting the box Dovecot is failing to start. When starting manually I get this message:

Starting IMAP/POP3 mail server: dovecotLast died with error (see error log for more information): Time just moved backwards by 15 seconds. This might cause a lot of problems, so I'll just kill myself now. http://wiki.dovecot.org/TimeMovedBackwards

Relevant part of syslog:

Jul 8 19:10:42 serverb3 bubba-networkmanager: Server timed out, terminating
Jul 8 19:10:42 serverb3 bubba-networkmanager: Daemon terminating
Jul 8 19:10:42 serverb3 bubba-networkmanager: Shutting down
Jul 8 19:10:28 serverb3 dovecot: Fatal: Time just moved backwards by 15 seconds. This might cause a lot of problems, so I'll just kill myself now. http://wiki.dovecot.org/TimeMovedBackwards

Not sure why it's changing. I think the bubba-network manager entry is a red-herring. Searching for that on these forums (using Google, the forums' own search facility won't let me) shows that people have been getting these errors since forever with no apparent problem.

So what's changing the time by 15 seconds? Investigations continue...

This is an issue that we've encountered before, and there should be an solution running that will resztart dovecot if it is terminated due to time revesal issue, though it might be only in certain situations that failsafe is present, for example when changing time setting.

Hi Carl, have you read through the wiki page that is mentioned in the error? There are a couple of potential explanations and fixes there.

[RandomUsername]

[Bug 5]
Samba daemon will not start if the wifi is enabled. However, starting samba first then enabling the wifi works fine.

The only thing I can see in the syslog is this:

Jul 9 13:16:53 serverb3 smbd[26120]: [2011/07/09 13:16:53.474838, 0] param/loadparm.c:7588(lp_do_parameter)
Jul 9 13:16:53 serverb3 smbd[26120]: Ignoring unknown parameter "interface"
Jul 9 13:16:53 serverb3 smbd[26120]: [2011/07/09 13:16:53.490481, 0] param/loadparm.c:7588(lp_do_parameter)
Jul 9 13:16:53 serverb3 smbd[26120]: Ignoring unknown parameter "interface"

This is from samba/log.all:

[2011/07/09 13:16:53.490278, 1] param/loadparm.c:6494(map_parameter)
Unknown parameter encountered: "interface"
[2011/07/09 13:16:53.490481, 0] param/loadparm.c:7588(lp_do_parameter)
Ignoring unknown parameter "interface"
[2011/07/09 13:16:53.622680, 0] lib/interface.c:542(load_interfaces)
WARNING: no network interfaces found
[2011/07/09 13:16:53.757681, 0] smbd/server.c:659(open_sockets_smbd)
open_sockets_smbd: No sockets available to bind to.
[2011/07/09 13:16:53.758655, 0] smbd/server.c:894(exit_server_common)
===============================================================
[2011/07/09 13:16:53.758832, 0] smbd/server.c:896(exit_server_common)
Abnormal server exit: open_sockets_smbd() failed
[2011/07/09 13:16:53.758936, 0] smbd/server.c:897(exit_server_common)
===============================================================
[2011/07/09 13:16:53.759666, 0] lib/util.c:1569(log_stack_trace)
BACKTRACE: 0 stack frames:
[2011/07/09 13:16:53.759871, 0] lib/fault.c:326(dump_core)
dumping core in /var/log/samba/cores/smbd

[EDIT]
OK, I've worked out what's happening here. When the wifi is enabled or disabled from the web interface the script is editing smb.conf but the script must be buggy (buggie?). I guess it's supposed to change the parameter "interfaces = eth1" to "interfaces = br0". However, it's adding "interface = br0" - note the missing "s" which is what is being complained about in the samba log. Also, it's leaving any previous "interfaces" parameter as well so I've got both these lines in smb.conf:

Code: Select all

interfaces = eth1
interface = br0

Even with the typo fixed I'm not sure you can have two lines the same. The correct formatting would be:

Code: Select all

interfaces = eth1 br0

However, just

Code: Select all

interfaces = br0

works fine.

Have been digging around in the code, and sadly, I've been unable to indicate anything that would set to "interface" instead of "interfaces"; I'll dig further, but can't guarantee I will find it in our code.

Maybe it's to do with my slightly modified smb.conf - I've added a couple of extra shares and change the permissions on the default ones. When I get home I will post my full smb.conf.

[DanielM]

So what's changing the time by 15 seconds? Investigations continue...

This is an issue that we've encountered before, and there should be an solution running that will resztart dovecot if it is terminated due to time revesal issue, though it might be only in certain situations that failsafe is present, for example when changing time setting.

The occasions that this has happened to me was after the B3 had crashed (which happened quite a lot with the old kernel, never happened after upgrading to 2.6.38), I guess things get kinda strange when dovecot is being terminated that way...

/Daniel

[carl]

[Bug 3]
Was just checking output of "dhclient" prior to setting DNS servers to be those of OpenDNS.
dhclient gives this output (note the syntax errors reported in //etc/dhcp/dhclient-exit-hooks.d/samba):

Code: Select all

Easyfind: nothing todo.
Enter Fallback (PREINIT)
Shutting down dnsmasq
Stopping DNS forwarder and DHCP server: dnsmasq.
Exit Fallback (PREINIT)
Firwall rewite, nothing todo.
/etc/dhcp/dhclient-exit-hooks.d/samba: line 78: syntax error near unexpected token `;'
/etc/dhcp/dhclient-exit-hooks.d/samba: line 78: `	if [ $reason != "PREINIT" ] && [ $reason != "RELEASE" ] && [ -x /etc/rc2.d/S??samba ] &&; then'
Easyfind: nothing todo.
Firwall rewite, nothing todo.
[b]/etc/dhcp/dhclient-exit-hooks.d/samba: line 78: syntax error near unexpected token `;'[/b][i][/i]
/etc/dhcp/dhclient-exit-hooks.d/samba: line 78: `	if [ $reason != "PREINIT" ] && [ $reason != "RELEASE" ] && [ -x /etc/rc2.d/S??samba ] &&; then'
Reloading /etc/samba/smb.conf: smbd only.
Launching easyfind update script
{"opcode":1,"record":{"domain_id":"1","content":"188.220.146.186","name":"darren.myownb3.com","type":"A","id":"541"},"error":false,"operation":"IPupdate"}
Update firewall config
Activate updated firewall config.
/etc/dhcp/dhclient-exit-hooks.d/samba: line 78: syntax error near unexpected token `;'
/etc/dhcp/dhclient-exit-hooks.d/samba: line 78: `	if [ $reason != "PREINIT" ] && [ $reason != "RELEASE" ] && [ -x /etc/rc2.d/S??samba ] &&; then'

Strangely, nano reports that this file only has 76 lines so I'm not sure how there could be an error on line 78

Fxied, thanks for the headsup; Was a missing commit before when I fixed an issue with interrupted samba transfers due to dhcp renews, strangly an extra "&&" had been snuggling into the repo.

[DanielM]

Todays upgrade gave me an error that I don't know exactly what to do about:

Code: Select all

Läs:1 http://b3.update.excito.org/ hugo/main bubba-backend all 2.3~rc1 [124 kB]
Läs:2 http://b3.update.excito.org/ hugo/main libphp-codeigniter all 1.7.2-5 [289 kB]
Läs:3 http://b3.update.excito.org/ hugo/main bubba-album all 2.3~rc2 [4 949 kB]
Läs:4 http://b3.update.excito.org/ hugo/main bubba-frontend all 2.3~rc4 [888 kB]
Hämtade 6 251 kB på 1s (5 924 kB/s)
(Läser databasen ... 47008 filer och kataloger installerade.)
Förbereder att ersätta bubba-backend 2.2.29 (med .../bubba-backend_2.3~rc1_all.deb) ...
Saving firewall...done.
Usage: /etc/init.d/smbrestart {start|stop|restart|force-reload}
invoke-rc.d: initscript smbrestart, action "stop" failed.
dpkg: varning: underprocess gammalt pre-removal-skript gav felkod 3
dpkg - försöker skript från det nya paketet istället ...
Saving firewall...done.
Usage: /etc/init.d/smbrestart {start|stop|restart|force-reload}
invoke-rc.d: initscript smbrestart, action "stop" failed.
dpkg: fel vid hantering av /var/cache/apt/archives/bubba-backend_2.3~rc1_all.deb (--unpack):
 underprocess nytt pre-removal-skript gav felkod 3
configured to not write apport reports
                                      Setting up firewall...done.
Restarting samba...Stopping Samba daemons: nmbd smbd.
Starting Samba daemons: nmbd smbd.
done.
Förbereder att ersätta libphp-codeigniter 1.7.2-4 (med .../libphp-codeigniter_1.7.2-5_all.deb) ...
Packar upp ersättande libphp-codeigniter ...
Förbereder att ersätta bubba-album 2.3~rc1 (med .../bubba-album_2.3~rc2_all.deb) ...
Packar upp ersättande bubba-album ...
Förbereder att ersätta bubba-frontend 2.3~rc2 (med .../bubba-frontend_2.3~rc4_all.deb) ...
Packar upp ersättande bubba-frontend ...
Fel uppstod vid hantering:
 /var/cache/apt/archives/bubba-backend_2.3~rc1_all.deb
E: Sub-process /usr/bin/dpkg returned an error code (1)

/Daniel

[RandomUsername]

Regarding the photo album - the PHP errors are fixed with the recent update but the management tools still don't seem to be available. I've uploaded a video to demonstrate. I'm just trying to use my normal account, not admin.

Regarding the wifi disable/enable mauling smb.conf and /etc/hosts - I put my smb.conf back to default and I'm finding this problem to be extremely reproducible.

For some reason, my smb.conf is having all comments removed but this is what it looks like after turning wifi off and then on again:

Code: Select all

; Config::Simple 4.59
; Tue Jul 12 23:10:04 2011

[printers]
guest ok=yes
comment=Attached printers
browseable=no
printable=yes
path=/var/spool/samba
use client driver=yes

[global]
log file=/var/log/samba/log.all
guest account=nobody
load printers=yes
obey pam restrictions=yes
socket options=TCP_NODELAY
interfaces=eth1
map to guest=Bad User
show add printer wizard=no
encrypt passwords=true
interface=br0
passdb backend=tdbsam
dns proxy=no
printing=cups
server string=%h Miniserver
invalid users=root
workgroup=excito
printcap name=cups
syslog=0
security=user
preferred master=yes
max log size=1000
disable spoolss=Yes
bind interfaces only=yes

[home]
hide dotfiles=yes
browseable=yes
writable=yes
path=/home
hide files=/lost+found/
guest ok=no
create mask=0764
directory mask=0775
comment=Home Directories
hide special files=yes
public=no
user=@users

[storage]
writable=yes
path=/home/storage
force directory mode=0777
force group=users
hide files=/lost+found/
guest ok=yes
force create mode=0777
comment=Common storage
hide special files=yes
public=yes

Also, just in case it was lost in the noise earlier - Carl, the dovecot not starting issue seems to be well explained in the wiki page the error points to. It also has a few potential fixes.

Darren.

[carl]

Regarding the photo album - the PHP errors are fixed with the recent update but the management tools still don't seem to be available. I've uploaded a video to demonstrate. I'm just trying to use my normal account, not admin.

Regarding the wifi disable/enable mauling smb.conf and /etc/hosts - I put my smb.conf back to default and I'm finding this problem to be extremely reproducible.

For some reason, my smb.conf is having all comments removed but this is what it looks like after turning wifi off and then on again:

Code: Select all

; Config::Simple 4.59
; Tue Jul 12 23:10:04 2011

[printers]
guest ok=yes
comment=Attached printers
browseable=no
printable=yes
path=/var/spool/samba
use client driver=yes

[global]
log file=/var/log/samba/log.all
guest account=nobody
load printers=yes
obey pam restrictions=yes
socket options=TCP_NODELAY
interfaces=eth1
map to guest=Bad User
show add printer wizard=no
encrypt passwords=true
interface=br0
passdb backend=tdbsam
dns proxy=no
printing=cups
server string=%h Miniserver
invalid users=root
workgroup=excito
printcap name=cups
syslog=0
security=user
preferred master=yes
max log size=1000
disable spoolss=Yes
bind interfaces only=yes

[home]
hide dotfiles=yes
browseable=yes
writable=yes
path=/home
hide files=/lost+found/
guest ok=no
create mask=0764
directory mask=0775
comment=Home Directories
hide special files=yes
public=no
user=@users

[storage]
writable=yes
path=/home/storage
force directory mode=0777
force group=users
hide files=/lost+found/
guest ok=yes
force create mode=0777
comment=Common storage
hide special files=yes
public=yes

Also, just in case it was lost in the noise earlier - Carl, the dovecot not starting issue seems to be well explained in the wiki page the error points to. It also has a few potential fixes.

Darren.

Thanks for the video; I must say I'm speechless, I can't think of anything that could result in an behavior like that. The only thing straight of the head I could think of debugging is either if you've had some JavaScript errors (firebug is your friend), and perhaps, but improbable, if there exists an user "darren" in /etc/bubba-auth.xml. Also, could you check what the server returns when you are trying to login (firebug again to rescue) (You can tail /var/log/apache/error.log as well to see if something out of the ordinary pops up).

For the smb issue, I would need to dig deeper into that; Could you point out the exact minimal sequence to reproduce the behavior? i.e. which pages you are accessing and what actions you are doing, and/or if you are rebooting or not.

Regarding dovecot, I notice now that there has been added some solutions to that page since I visited it before. I'm gonna look into applying a fix or two to reduce the impact of the time reduction problem.

[RandomUsername]

Hi Carl,

First thing I did was to do

Code: Select all

aptitude reinstall bubba-album

just to make sure. It installed version 2.3~rc3.

The first time I browsed to the photo albums I got the php errors again so maybe you haven't hidden them like you thought :S

Screenshot-B3 - photo album - Mozilla Firefox-1.png (155.01 KiB) Viewed 16834 times

When trying to log into the photo album site these errors are coming up in /var/log/apache2/error.log:

Code: Select all

[Wed Jul 13 15:55:52 2011] [error] [client 192.168.10.98] File does not exist: /usr/share/album/views/_css/images, referer: https://serverb3/album/views/_css/jquery.ui.theme.default.css
[Wed Jul 13 15:55:55 2011] [error] [client 192.168.10.98] File does not exist: /usr/share/album/views/_css/images, referer: https://serverb3/album/views/_css/jquery.ui.theme.default.css

Not sure if that's relevant, but the directory /usr/share/album/views/_css/images definitely doesn't exist.

This is the first time I've used Firebug so I don't know what output you want. I didn't see any errors but it did log this failed POST attempt when trying to log in:

Code: Select all

{"success":false,"has_access":false}

This is the contents of /etc/bubba_auth.xml:

Code: Select all

<?xml version="1.0" encoding="utf-8"?>
<auth>
</auth>

Should the admin account not be listed there?

Regarding smb.conf and /etc/hosts, literally all I did was go to https://serverb3/admin/network/wlan, turn the wireless access point either on or off (depending on the initial status) and click "update" and my hostname is removed from /etc/hosts and smb.conf gets the invalid parameter added. I can reproduce this every time. Is there anything I can do to track down the culprit?

[carl]

Hi Carl,

First thing I did was to do

Code: Select all

aptitude reinstall bubba-album

just to make sure. It installed version 2.3~rc3.

The first time I browsed to the photo albums I got the php errors again so maybe you haven't hidden them like you thought :S

Screenshot-B3 - photo album - Mozilla Firefox-1.png

When trying to log into the photo album site these errors are coming up in /var/log/apache2/error.log:

Code: Select all

[Wed Jul 13 15:55:52 2011] [error] [client 192.168.10.98] File does not exist: /usr/share/album/views/_css/images, referer: https://serverb3/album/views/_css/jquery.ui.theme.default.css
[Wed Jul 13 15:55:55 2011] [error] [client 192.168.10.98] File does not exist: /usr/share/album/views/_css/images, referer: https://serverb3/album/views/_css/jquery.ui.theme.default.css

Not sure if that's relevant, but the directory /usr/share/album/views/_css/images definitely doesn't exist.

This is the first time I've used Firebug so I don't know what output you want. I didn't see any errors but it did log this failed POST attempt when trying to log in:

Code: Select all

{"success":false,"has_access":false}

This is the contents of /etc/bubba_auth.xml:

Code: Select all

<?xml version="1.0" encoding="utf-8"?>
<auth>
</auth>

Should the admin account not be listed there?

Regarding smb.conf and /etc/hosts, literally all I did was go to https://serverb3/admin/network/wlan, turn the wireless access point either on or off (depending on the initial status) and click "update" and my hostname is removed from /etc/hosts and smb.conf gets the invalid parameter added. I can reproduce this every time. Is there anything I can do to track down the culprit?

The version I assume you've installed is 2.3~rc2 if bubba-album
But nevertheless, the apache log lines can be ignored, it's only because it's the same css as used my admin panel. I can't really deduce what might be wrong here, haven't been able to reproduce the album situation at all. Would it be possible for you to grant me access to investigate on your box what might have happened?

The POST reply indicate that the server have managed to properly fail; is it possible for you to post here the query as well?

The bubba-auth.xml is an ACL outside the normal auth system, that I made primarly for the album, but is designed to be usable for other tasks as well; It the moment it's only used to contain album users.

But... The smb.conf and /etc/hosts issue I was able to reproduce, and will fix it asap.

[DanielM]

Just like to note that bubba-backend still fails to install for me. I get the same error in rc2 as in rc1:

Code: Select all

dpkg: fel vid hantering av /var/cache/apt/archives/bubba-backend_2.3~rc2_all.deb            (--unpack):
 underprocess nytt pre-removal-skript gav felkod 3
configured to not write apport reports

/Daniel

[RandomUsername]

Hey Carl,

Excuse the typo - you are correct, it should be 2.3~RC2.

Good news about smb.conf and hosts.

The POST reply indicate that the server have managed to properly fail; is it possible for you to post here the query as well?

Sorry, not 100% sure what you want, could you point me to what options to enable in FireBug?

I can see my username and password being sent in the POST request and they're both correct.

If it helps, this is the header information for the POST request:

Code: Select all

Date	Wed, 13 Jul 2011 20:19:50 GMT
Server	Apache/2.2.16 (Debian)
X-Powered-By	PHP/5.3.3-7+squeeze3
Content-Length	36
Keep-Alive	timeout=15, max=97
Connection	Keep-Alive
Content-Type	application/json
Request Headers
Host	sharefamily.co.cc
User-Agent	Mozilla/5.0 (Windows NT 5.1; rv:5.0) Gecko/20100101 Firefox/5.0
Accept	application/json, text/javascript, */*
Accept-Language	en-gb
Accept-Encoding	gzip, deflate
Accept-Charset	ISO-8859-1,utf-8;q=0.7,*;q=0.7
DNT	1
Connection	keep-alive
Content-Type	application/x-www-form-urlencoded; charset=UTF-8
X-Requested-With	XMLHttpRequest
Referer	https://sharefamily.co.cc/album/
Content-Length	37
Cookie	template=jstyle_blue; album_session=a%3A4%3A%7Bs%3A10%3A%22session_id%22%3Bs%3A32%3A%22b9d48fbf96e9d94c719fd55be97cde7c%22%3Bs%3A10%3A%22ip_address%22%3Bs%3A13%3A%2288.217.168.15%22%3Bs%3A10%3A%22user_agent%22%3Bs%3A50%3A%22Mozilla%2F5.0+%28Windows+NT+5.1%3B+rv%3A5.0%29+Gecko%2F2010010%22%3Bs%3A13%3A%22last_activity%22%3Bs%3A10%3A%221310588386%22%3B%7D008ba4ba32d2a271e05ebee998259f87; PHPSESSID=l9tpst2auj6o605h31dti1n5v5; roundcube_sessid=0ub6n3i8ksqmgb0eueq7fpamq6

Regarding access to my box, that's not a problem but would have to wait until I'm home tomorrow. Do you want root / admin access?

Thanks.

Darren.

[Nrde]

I just wanted to add that I have the exact same problem as Darren.

Also I see one album (that I created with the release FW) when I type the album address to to the location field of my browser. If I click the photos icon on the admin interface I can't see the album.

When I try to log in while in album the /album/index.php/login/json return the same success, false json as Darren (with corect username & pw).

Is there some hardcoded ports in some of the relevant scripts as I'm using my webinterface with port 8080?

And it seems that I can't see even the initial album if I logged in from the "main" page. (not by clicking the icon or typing the address to location field.)

[carl]

...

Regarding access to my box, that's not a problem but would have to wait until I'm home tomorrow. Do you want root / admin access?

Thanks.

Darren.

Yes, I would need root access, and also possibility to access the web interface.