[Bug 1325683] [NEW] Having the string '$pid' in the config file breaks the init script
Public bug reported:
I have the following line in the main config file
(/etc/nginx/nginx.conf):
log_format syslog 'nginx[$pid]: $host $remote_addr $remote_user
$request $status $body_bytes_sent $http_referer $http_user_agent';
This line alone makes it impossible to start nginx. The init script
parses the pid file path from the config file, but does that in a
unreliable way:
PID=$(awk -F'[ \t;]+' '/[^#]pid/ {print $2}' /etc/nginx/nginx.conf)
For my config file, this returns two lines, the first one with the pid
file, the second line with just 'log_format':
$ awk -F'[ \t;]+' '/[^#]pid/ {print $2}' /etc/nginx/nginx.conf
/var/run/nginx.pid
log_format
$
Later on when the init script attempts to start nginx, it fails:
$ start-stop-daemon --start --quiet --pidfile /var/run/nginx.pid log_format
--exec /usr/sbin/nginx --
nginx: invalid option: log_format
Ubuntu 14.04, everything updated today.
** Affects: nginx (Ubuntu)
Importance: Undecided
Status: New
--
You received this bug notification because you are a member of Ubuntu
Server Team, which is subscribed to nginx in Ubuntu.
https://bugs.launchpad.net/bugs/1325683
Title:
Having the string '$pid' in the config file breaks the init script
To manage notifications about this bug go to:
https://bugs.launchpad.net/ubuntu/+source/nginx/+bug/1325683/+subscriptions
--
Ubuntu-server-bugs mailing list
Ubuntu-server-bugs@lists.ubuntu.com
Modify settings or unsubscribe at:
https://lists.ubuntu.com/mailman/listinfo/ubuntu-server-bugs
[Bug 1325683] [NEW] Having the string '$pid' in the config file breaks the init script
Public bug reported:
I have the following line in the main config file
(/etc/nginx/nginx.conf):
log_format syslog 'nginx[$pid]: $host $remote_addr $remote_user
$request $status $body_bytes_sent $http_referer $http_user_agent';
This line alone makes it impossible to start nginx. The init script
parses the pid file path from the config file, but does that in a
unreliable way:
PID=$(awk -F'[ \t;]+' '/[^#]pid/ {print $2}' /etc/nginx/nginx.conf)
For my config file, this returns two lines, the first one with the pid
file, the second line with just 'log_format':
$ awk -F'[ \t;]+' '/[^#]pid/ {print $2}' /etc/nginx/nginx.conf
/var/run/nginx.pid
log_format
$
Later on when the init script attempts to start nginx, it fails:
$ start-stop-daemon --start --quiet --pidfile /var/run/nginx.pid log_format
--exec /usr/sbin/nginx --
nginx: invalid option: log_format
Ubuntu 14.04, everything updated today.
** Affects: nginx (Ubuntu)
Importance: Undecided
Status: New
--
You received this bug notification because you are a member of Ubuntu
Bugs, which is subscribed to Ubuntu.
https://bugs.launchpad.net/bugs/1325683
Title:
Having the string '$pid' in the config file breaks the init script
To manage notifications about this bug go to:
https://bugs.launchpad.net/ubuntu/+source/nginx/+bug/1325683/+subscriptions
--
ubuntu-bugs mailing list
ubuntu-bugs@lists.ubuntu.com
https://lists.ubuntu.com/mailman/listinfo/ubuntu-bugs
[Bug 380663] Re: open-iscsi initiator tries to resolve ipv6 address of target and fails
This is not a bug in OpenSolaris, it is a problem in open-iscsi. The OpenSolaris iscsi implementation sends all IP addresses through which hosts can connect to it. In my case these are three addresses: IPv4, IPv6 link local and IPv6 global: iscsiadm: finished reading text PDU, 48 hdr, 0 ah, 591 data, 1 pad iscsiadm: TargetName=iqn.1986-03.com.sun:02:xxx iscsiadm: TargetAddress=192.168.0.80,1 iscsiadm: TargetAddress=[fe80::230:1bff:fe12:3456],1 iscsiadm: TargetAddress=[2001:1234:5678:0:230:1bff:fe12:3456],1 The problem is that the second address, the IPv6 link local one, can only be used to connect to the host when the application supplies the 'zone index' to the kernel (interface which the kernel should use). This is a property of IPv6 link local addresses. Open-iscsi doesn't support this logic yet, but when it encounters such address it skips the whole target and goes to the next one. This even if there are other addresses which would work fine. The attached patch fixes it by not failing when encountering an unknown target address. When it does it skips the address and tries the next one. With this patch I can connect to targets on OpenSolaris just fine. ** Attachment added: open-iscsi-try-all-target-addresses.patch http://launchpadlibrarian.net/29488027/open-iscsi-try-all-target-addresses.patch ** Changed in: open-iscsi (Ubuntu) Status: Invalid = New -- open-iscsi initiator tries to resolve ipv6 address of target and fails https://bugs.launchpad.net/bugs/380663 You received this bug notification because you are a member of Ubuntu Server Team, which is subscribed to open-iscsi in ubuntu. -- Ubuntu-server-bugs mailing list Ubuntu-server-bugs@lists.ubuntu.com Modify settings or unsubscribe at: https://lists.ubuntu.com/mailman/listinfo/ubuntu-server-bugs
[Bug 380663] Re: open-iscsi initiator tries to resolve ipv6 address of target and fails
This is not a bug in OpenSolaris, it is a problem in open-iscsi. The OpenSolaris iscsi implementation sends all IP addresses through which hosts can connect to it. In my case these are three addresses: IPv4, IPv6 link local and IPv6 global: iscsiadm: finished reading text PDU, 48 hdr, 0 ah, 591 data, 1 pad iscsiadm: TargetName=iqn.1986-03.com.sun:02:xxx iscsiadm: TargetAddress=192.168.0.80,1 iscsiadm: TargetAddress=[fe80::230:1bff:fe12:3456],1 iscsiadm: TargetAddress=[2001:1234:5678:0:230:1bff:fe12:3456],1 The problem is that the second address, the IPv6 link local one, can only be used to connect to the host when the application supplies the 'zone index' to the kernel (interface which the kernel should use). This is a property of IPv6 link local addresses. Open-iscsi doesn't support this logic yet, but when it encounters such address it skips the whole target and goes to the next one. This even if there are other addresses which would work fine. The attached patch fixes it by not failing when encountering an unknown target address. When it does it skips the address and tries the next one. With this patch I can connect to targets on OpenSolaris just fine. ** Attachment added: open-iscsi-try-all-target-addresses.patch http://launchpadlibrarian.net/29488027/open-iscsi-try-all-target-addresses.patch ** Changed in: open-iscsi (Ubuntu) Status: Invalid = New -- open-iscsi initiator tries to resolve ipv6 address of target and fails https://bugs.launchpad.net/bugs/380663 You received this bug notification because you are a member of Ubuntu Bugs, which is subscribed to Ubuntu. -- ubuntu-bugs mailing list ubuntu-bugs@lists.ubuntu.com https://lists.ubuntu.com/mailman/listinfo/ubuntu-bugs
[Bug 315455] Re: graphviz cairo rasterizer doesn't correctly draw rounded objects
** Attachment added: dot graph with rounded corners http://launchpadlibrarian.net/21054229/test.dot -- graphviz cairo rasterizer doesn't correctly draw rounded objects https://bugs.launchpad.net/bugs/315455 You received this bug notification because you are a member of Ubuntu Bugs, which is subscribed to Ubuntu. -- ubuntu-bugs mailing list ubuntu-bugs@lists.ubuntu.com https://lists.ubuntu.com/mailman/listinfo/ubuntu-bugs
[Bug 315455] [NEW] graphviz cairo rasterizer doesn't correctly draw rounded objects
Public bug reported: Ubuntu 8.04.1, graphviz-2.16-3ubuntu2 When rasterizing dot graphs that contain rounded objects (such as the test graph in the attachement: test.dot), the rounded corners aren't being drawn correctly. This only affects the cairo rasterizer (dot -Tpng:cairo:cairo), the gd rasterizer works fine (dot -Tpng:gd:gd). I attached the rasterized graph for your convenience. ** Affects: graphviz-cairo (Ubuntu) Importance: Undecided Status: New -- graphviz cairo rasterizer doesn't correctly draw rounded objects https://bugs.launchpad.net/bugs/315455 You received this bug notification because you are a member of Ubuntu Bugs, which is subscribed to Ubuntu. -- ubuntu-bugs mailing list ubuntu-bugs@lists.ubuntu.com https://lists.ubuntu.com/mailman/listinfo/ubuntu-bugs
[Bug 315455] Re: graphviz cairo rasterizer doesn't correctly draw rounded objects
** Attachment added: Rasterized dot graph, using -Tpng:cairo:cairo http://launchpadlibrarian.net/21054260/dot-graph.png -- graphviz cairo rasterizer doesn't correctly draw rounded objects https://bugs.launchpad.net/bugs/315455 You received this bug notification because you are a member of Ubuntu Bugs, which is subscribed to Ubuntu. -- ubuntu-bugs mailing list ubuntu-bugs@lists.ubuntu.com https://lists.ubuntu.com/mailman/listinfo/ubuntu-bugs
[Bug 120608] Add /usr/local/lib to ld search path
Public bug reported: Packages usually have prefix=/usr/local unless the user overrides it (./configure --prefix=???). If Ubuntu users follow the simple './configure make sudo make install' guide packages will install libraries into /usr/local/lib. But since ld doesn't search that directory, applications fail to start or work. I'm a software developer and everytime someone tells me the application doesn't work on his Ubuntu/Debian box, I can immediately assume that it's due to the above problem. It's not just my package that's broken (with the default settings) because of that, it's all packages that are out there. So it would make much more sense to add /usr/local/lib to the ld search path than telling Ubuntu/Debian users to do it by hand or use --prefix=/usr. ** Affects: Ubuntu Importance: Undecided Status: Unconfirmed -- Add /usr/local/lib to ld search path https://bugs.launchpad.net/bugs/120608 You received this bug notification because you are a member of Ubuntu Bugs, which is the bug contact for Ubuntu. -- ubuntu-bugs mailing list ubuntu-bugs@lists.ubuntu.com https://lists.ubuntu.com/mailman/listinfo/ubuntu-bugs
[Bug 117131] 32bit development libraries missing for amd64 systems systems
Public bug reported: At least /usr/lib32/libGL.so and /usr/lib32/libX11.so are not available on amd64 systems, making it impossible to compile/build software from source without having the user create these symlinks themselves. Maybe you should create a new package called ia32-libs-dev? Note that I'm not running ubuntu, I'm just a software developer. My software doesn't have an ubuntu package yet and the users have to build it from source. Being able to build the 32bit version on amd64 systems is very important! Besides of the missing development libraries, which is quite easy to fix by the users themselves, a much more severe problem is that the 32bit version of libXv.so.1 is missing, but there's already a bug report for that, #53535. ** Affects: Ubuntu Importance: Undecided Status: Unconfirmed -- 32bit development libraries missing for amd64 systems systems https://bugs.launchpad.net/bugs/117131 You received this bug notification because you are a member of Ubuntu Bugs, which is the bug contact for Ubuntu. -- ubuntu-bugs mailing list ubuntu-bugs@lists.ubuntu.com https://lists.ubuntu.com/mailman/listinfo/ubuntu-bugs
[Bug 117131] Re: 32bit development libraries missing for amd64 systems systems
It compiles just fine on 64bit, but the users _need_ to build it as 32bit software in order to make it work with wine. This is the project: http://www.neopsis.com/projects/yukon/ -- 32bit development libraries missing for amd64 systems systems https://bugs.launchpad.net/bugs/117131 You received this bug notification because you are a member of Ubuntu Bugs, which is the bug contact for Ubuntu. -- ubuntu-bugs mailing list ubuntu-bugs@lists.ubuntu.com https://lists.ubuntu.com/mailman/listinfo/ubuntu-bugs
[Bug 117131] Re: 32bit development libraries missing for amd64 systems systems
.. see also ticket 8 (http://www.neopsis.com/projects/yukon/ticket/8) -- 32bit development libraries missing for amd64 systems systems https://bugs.launchpad.net/bugs/117131 You received this bug notification because you are a member of Ubuntu Bugs, which is the bug contact for Ubuntu. -- ubuntu-bugs mailing list ubuntu-bugs@lists.ubuntu.com https://lists.ubuntu.com/mailman/listinfo/ubuntu-bugs
Re: [Bug 117131] Re: 32bit development libraries missing for amd64 systems systems
Thanks for the further explanation. I have heard that there are plans underway to supply 64-bit wine for 64-bit architectures, and 32-bit wine for 32-bit architectures, which may address the needs of some of your users. Not always. If the user wants to run Windows games (which are usually 32bit) he'll need the 32bit version of wine even on amd64 systems. Most users are trying to capture World of Warcraft or other similar windows games so that's why they're trying to build the 32bit version of yukon. As I understand it, the current recommendation is to use a 32-bt environment to compile 32-bit applications, and then run them in the 64-bit environment. That's quite unfortunate. How should I explain my users to install the package on amd64 systems? Not everybody has access to a 32bit ubuntu box. I've marked this bug confirmed, as it is clear that there is not an ia32 -libs-dev on amd64, and that it doesn't contain libXv.so.1, but it may still be rejected upon review by the developers, as there may be reasons why beyond those my research discovered. The 32bit libXv.so.1 should be part of ia32-libs as indicated by the other bug report, libXv.so (along with libX11.so and other 32bit development libraries) should be part of ia32-libs-dev (if you indeed decide to create that package). -- 32bit development libraries missing for amd64 systems systems https://bugs.launchpad.net/bugs/117131 You received this bug notification because you are a member of Ubuntu Bugs, which is the bug contact for Ubuntu. -- ubuntu-bugs mailing list ubuntu-bugs@lists.ubuntu.com https://lists.ubuntu.com/mailman/listinfo/ubuntu-bugs
[Bug 117236] new package: seom
Public bug reported: The 'seom' library (http://www.neopsis.com/projects/yukon/) is used by 'yukon' (http://www.neopsis.com/projects/yukon/). Building the package should be straight-forward, eg. ./configure; make; sudo make install. tarballs can be found here: http://dbservice.com/ftpdir/tom/seom/trunk/ ** Affects: Ubuntu Importance: Undecided Status: Unconfirmed ** Tags: needs-packaging -- new package: seom https://bugs.launchpad.net/bugs/117236 You received this bug notification because you are a member of Ubuntu Bugs, which is the bug contact for Ubuntu. -- ubuntu-bugs mailing list ubuntu-bugs@lists.ubuntu.com https://lists.ubuntu.com/mailman/listinfo/ubuntu-bugs
[Bug 117236] Re: new package: seom
.. of course the seom homepage is here: http://www.neopsis.com/projects/seom/ -- new package: seom https://bugs.launchpad.net/bugs/117236 You received this bug notification because you are a member of Ubuntu Bugs, which is the bug contact for Ubuntu. -- ubuntu-bugs mailing list ubuntu-bugs@lists.ubuntu.com https://lists.ubuntu.com/mailman/listinfo/ubuntu-bugs
[Bug 117235] new package: seom
Public bug reported: The 'seom' library is used by 'yukon' (http://www.neopsis.com/projects/yukon/). Building the package should be straight-forward, eg. ./configure; make; sudo make install. ** Affects: Ubuntu Importance: Undecided Status: Unconfirmed ** Tags: needs-packaging -- new package: seom https://bugs.launchpad.net/bugs/117235 You received this bug notification because you are a member of Ubuntu Bugs, which is the bug contact for Ubuntu. -- ubuntu-bugs mailing list ubuntu-bugs@lists.ubuntu.com https://lists.ubuntu.com/mailman/listinfo/ubuntu-bugs
