[Bug 1810535] Re: go 1.10 fails to symlink misc into GOROOT
Same issue on 18.04. Resolved with $ cd /usr/lib/go-1.10/ ; sudo ln -s ../../share/go-1.10/misc misc $ lsb_release -rd Description:Ubuntu 18.04.3 LTS Release:18.04 $ apt-cache policy golang-1.10-go golang-1.10-go: Installed: 1.10.4-2ubuntu1~18.04.1 Candidate: 1.10.4-2ubuntu1~18.04.1 Version table: *** 1.10.4-2ubuntu1~18.04.1 500 500 http://us.archive.ubuntu.com/ubuntu bionic-updates/main amd64 Packages 100 /var/lib/dpkg/status 1.10.1-1ubuntu2 500 500 http://us.archive.ubuntu.com/ubuntu bionic/main amd64 Packages ** Tags added: bionic -- You received this bug notification because you are a member of Ubuntu Bugs, which is subscribed to Ubuntu. https://bugs.launchpad.net/bugs/1810535 Title: go 1.10 fails to symlink misc into GOROOT To manage notifications about this bug go to: https://bugs.launchpad.net/ubuntu/+source/golang-1.10/+bug/1810535/+subscriptions -- ubuntu-bugs mailing list ubuntu-bugs@lists.ubuntu.com https://lists.ubuntu.com/mailman/listinfo/ubuntu-bugs
[Bug 1331320] Re: Differences between test-kickseed and real execution
** Branch linked: lp:~hamiltont-x/kickseed/kickseed -- You received this bug notification because you are a member of Ubuntu Bugs, which is subscribed to Ubuntu. https://bugs.launchpad.net/bugs/1331320 Title: Differences between test-kickseed and real execution To manage notifications about this bug go to: https://bugs.launchpad.net/ubuntu/+source/kickseed/+bug/1331320/+subscriptions -- ubuntu-bugs mailing list ubuntu-bugs@lists.ubuntu.com https://lists.ubuntu.com/mailman/listinfo/ubuntu-bugs
[Bug 1331814] Re: Warning messages for handlers are never called
Also, I added a reference file for kickstart that lists all commands with all command options. By doing $ ./test-kickseed tests/full-full-06-2014.ks we can see which kickstart options are not currently supported, as they will print out LOG messages. It's not exactly test driven dev, but it's something ;-) -- You received this bug notification because you are a member of Ubuntu Bugs, which is subscribed to Ubuntu. https://bugs.launchpad.net/bugs/1331814 Title: Warning messages for handlers are never called To manage notifications about this bug go to: https://bugs.launchpad.net/ubuntu/+source/kickseed/+bug/1331814/+subscriptions -- ubuntu-bugs mailing list ubuntu-bugs@lists.ubuntu.com https://lists.ubuntu.com/mailman/listinfo/ubuntu-bugs
[Bug 1331814] Re: Warning messages for handlers are never called
Fixed. Ash/Dash don't support ERR traps, so I instead did this: set +e output=$(getopt -o '' -l $options -- $@) errout=$(getopt -o '' -l $options -- $@ 21 /dev/null) set -e if [[ -n $errout ]]; then warn_getopt $type warn $type Error: $errout fi This results in much better logging during install, without stopping output to console. For example: $ cat /var/log/syslog | grep kickseed ... snip ... Jun 19 01:56:26 kickseed: Failed to parse %pre options Jun 19 01:56:26 kickseed: %pre Error: getopt: unrecognized option '--log=/test.log' I've also integrated this with test-kickseed, which outputs this: $ ./test-kickseed custom-test.ks getopt: unrecognized option '--log=/test.log' LOG: Failed to parse some %pre options LOG: %pre Error: getopt: unrecognized option '--log=/test.log' As you can see, getopt still prints to stderr, but now we have log messages printed as well. ** Changed in: kickseed (Ubuntu) Status: New = Fix Committed ** Changed in: kickseed (Ubuntu) Assignee: (unassigned) = Hamilton Turner (hamiltont-x) -- You received this bug notification because you are a member of Ubuntu Bugs, which is subscribed to Ubuntu. https://bugs.launchpad.net/bugs/1331814 Title: Warning messages for handlers are never called To manage notifications about this bug go to: https://bugs.launchpad.net/ubuntu/+source/kickseed/+bug/1331814/+subscriptions -- ubuntu-bugs mailing list ubuntu-bugs@lists.ubuntu.com https://lists.ubuntu.com/mailman/listinfo/ubuntu-bugs
[Bug 976632] Re: nfs4 mounts are not mapping userids, uses nobody/nogroup
Just experienced this issue. Used TJ's fix with success, but I had to also restart nfs on the server - multiple restarts of idmapd did nothing. I did not try restarting the client. FWIW, here's my details. Primary difference is that my server is a CentOS box. Client: Linux 3 3.11.0-23-generic #40~precise1-Ubuntu SMP Wed Jun 4 22:06:36 UTC 2014 x86_64 x86_64 x86_64 GNU/Linux nfs-common 1:1.2.5-3ubuntu3.1 nfs4-acl-tools: 0.3.3-1 nfswatch: 4.99.11-1 Server: Linux snip 2.6.32-431.5.1.el6.x86_64 #1 SMP Wed Feb 12 00:41:43 UTC 2014 x86_64 x86_64 x86_64 GNU/Linux nfs-utils-1.2.3-39.el6.x86_64 nfs-utils-lib-1.1.5-6.el6.x86_64 nfs4-acl-tools-0.3.3-6.el6.x86_64 -- You received this bug notification because you are a member of Ubuntu Bugs, which is subscribed to Ubuntu. https://bugs.launchpad.net/bugs/976632 Title: nfs4 mounts are not mapping userids, uses nobody/nogroup To manage notifications about this bug go to: https://bugs.launchpad.net/ubuntu/+source/nfs-utils/+bug/976632/+subscriptions -- ubuntu-bugs mailing list ubuntu-bugs@lists.ubuntu.com https://lists.ubuntu.com/mailman/listinfo/ubuntu-bugs
[Bug 976632] Re: nfs4 mounts are not mapping userids, uses nobody/nogroup
Also, my client/server are sharing UIDs, which seems to imply this is independent of https://bugs.launchpad.net/ubuntu/+source/nfs- utils/+bug/966734 -- You received this bug notification because you are a member of Ubuntu Bugs, which is subscribed to Ubuntu. https://bugs.launchpad.net/bugs/976632 Title: nfs4 mounts are not mapping userids, uses nobody/nogroup To manage notifications about this bug go to: https://bugs.launchpad.net/ubuntu/+source/nfs-utils/+bug/976632/+subscriptions -- ubuntu-bugs mailing list ubuntu-bugs@lists.ubuntu.com https://lists.ubuntu.com/mailman/listinfo/ubuntu-bugs
[Bug 1331320] [NEW] Differences between test-kickseed and real execution
Public bug reported: I've encountered an odd issue, perhaps someone can help me isolate. When I run test-kickseed on the %pre block below, 'handler/pre.sh' reports a getopt failure parsing the %pre options **and continues executing**. (support for --log is the feature I'm currently adding into kickseed, the issue is that it continues executing) --- %pre --log=foobar.log echo should pass %end When I attempt an install with this %pre **kickseed halts**. This is because the call to getopt returns a non-zero exit code and errexit (set -e) is turned on. Halting seems to be the proper behavior, but test- kickseed continues executing. I originally thought it had something to do with my shell or my getopt implementation, but I wrote a small test.sh file (below) that shows that set -e works in a manner consistent with what I'm seeing during an actual install. This test file halts when running bash -e test.sh or sh -e test.sh, and test-kickseed does not. --test.sh--- #!/bin/sh -e bar () { set -o | grep errexit TEMP=`getopt -o '' -l foo: -- $@` echo good stuff } bar --foo=bar bar --log - $ ./test.sh errexit on good stuff errexit on getopt: unrecognized option '--log' -- So what's happening with test-kickseed? I did add a debug statement in handlers/pre.sh before the call to getopt, and errexit is definitely turned on. I added a debug statement after getopt, and the error code is 1. This seems like it should be halting, but for some insane reason it's not - with set -e I shouldn't even see the output telling me the error code, the program should terminate automatically. PS - I have moved all my getopt calls to their own lines, so this is not an issue with set -e and || ** Affects: kickseed (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/1331320 Title: Differences between test-kickseed and real execution To manage notifications about this bug go to: https://bugs.launchpad.net/ubuntu/+source/kickseed/+bug/1331320/+subscriptions -- ubuntu-bugs mailing list ubuntu-bugs@lists.ubuntu.com https://lists.ubuntu.com/mailman/listinfo/ubuntu-bugs
[Bug 1331320] Re: Differences between test-kickseed and real execution
I'm not insane, but I am wrong! This is indeed caused by set -e and ||. Specifically, the difference between how test-kickseed and initrd- kickseed call the kickseed function. test-kickseed calls: RET=0 (kickseed $1) || RET=$? This means that kickseed() is called as part of any command executed in a or || list, and therefore errexit has no effect anywhere inside of the kickseed function when it is called from test-kickseed. This is clearly intentional, but it also caused me substantial confusion due to different behavior, so I'll put in a tiny commit that causes a big warning message to be printed when $RET != 0. Currently the exit code is just being returned to the user, and it would take a particularly savvy user to check the exit code of test-kickseed every time and interpret that a non-zero means that the install will behave differently than the test. ** Changed in: kickseed (Ubuntu) Status: New = In Progress ** Changed in: kickseed (Ubuntu) Assignee: (unassigned) = Hamilton Turner (hamiltont-x) -- You received this bug notification because you are a member of Ubuntu Bugs, which is subscribed to Ubuntu. https://bugs.launchpad.net/bugs/1331320 Title: Differences between test-kickseed and real execution To manage notifications about this bug go to: https://bugs.launchpad.net/ubuntu/+source/kickseed/+bug/1331320/+subscriptions -- ubuntu-bugs mailing list ubuntu-bugs@lists.ubuntu.com https://lists.ubuntu.com/mailman/listinfo/ubuntu-bugs
[Bug 1331320] Re: Differences between test-kickseed and real execution
I believe http://bazaar.launchpad.net/~hamiltont-x/kickseed/kickseed/revision/298 addresses this well. I decided that using $RET was a poor choice, as it only checks the return of the kickseed function and can return a 0 value even if an internal function returns a non-zero value. I have instead disabled errexit when running test-kickseed and added a trap for catching ERR signals. This allows us to output a nice warning message that the installation will not work as expected because there was an ERR triggered while parsing the kickstart. Note: I used code from http://stackoverflow.com/q/6928946/119592, so if the licensing isn't ok then we may need to rewrite the ERR handler from scratch. If this is necessary someone let me know, it's only a few lines but the one I used is fairly comprehensive ** Changed in: kickseed (Ubuntu) Status: In Progress = Fix Committed -- You received this bug notification because you are a member of Ubuntu Bugs, which is subscribed to Ubuntu. https://bugs.launchpad.net/bugs/1331320 Title: Differences between test-kickseed and real execution To manage notifications about this bug go to: https://bugs.launchpad.net/ubuntu/+source/kickseed/+bug/1331320/+subscriptions -- ubuntu-bugs mailing list ubuntu-bugs@lists.ubuntu.com https://lists.ubuntu.com/mailman/listinfo/ubuntu-bugs
[Bug 1331814] [NEW] Warning messages for handlers are never called
Public bug reported: Multiple handlers use the following syntax: eval set -- $(getopt -o '' -l interpreter: -- $@) || { warn_getopt %pre; return; } As mentioned in /usr/share/doc/util-linux-ng-2.17.2/getopt-parse.bash, this will never call the warning: # We need TEMP as the `eval set --' would nuke the return value of getopt. TEMP=`getopt -o ab:c:: --long a-long,b-long:,c-long:: \ -n 'example.bash' -- $@` Here is a simple example to prove this: #!/bin/sh trap 'echo ERROR happened' ERR echo begin eval set -- $(false) echo end Because of this, almost every bug report mentions something along the lines of getopt printed an error. This is getout outputting to stderr, but the non-zero exit code is being swallowed and errexit is not happening. This coding was probably designed intentionally to disable automatic exit whenever getopt found an unexpected argument, due to the fact that this (currently) happens on almost every kickstart file as our syntax support is quite limited. However, it also hides the warning from the user, which I believe is the real bug. Moreover, the only error is printed briefly to a terminal and then never seen again, which makes it tough for people to put in bug reports. Perhaps a better solution would be to use set +x and set -e to temporarily disable errexit, and use a trap on ERR to ensure that these errors are caught and printed to both syslog and the current TTY. ** Affects: kickseed (Ubuntu) Importance: Undecided Status: New ** Description changed: Multiple handlers use the following syntax: - eval set -- $(getopt -o '' -l interpreter: -- $@) || { + eval set -- $(getopt -o '' -l interpreter: -- $@) || { warn_getopt %pre; return; } - As mentioned in /usr/share/doc/util-linux-ng-2.17.2/getopt-parse.bash, this will never - call the warning: + As mentioned in /usr/share/doc/util-linux-ng-2.17.2/getopt-parse.bash, this will never + call the warning: - # We need TEMP as the `eval set --' would nuke the return value of getopt. -TEMP=`getopt -o ab:c:: --long a-long,b-long:,c-long:: \ - -n 'example.bash' -- $@` + # We need TEMP as the `eval set --' would nuke the return value of getopt. + TEMP=`getopt -o ab:c:: --long a-long,b-long:,c-long:: \ + -n 'example.bash' -- $@` Here is a simple example to prove this: - #!/bin/sh - trap 'echo ERROR happened' ERR - echo begin - eval set -- $(false) - echo end + #!/bin/sh + trap 'echo ERROR happened' ERR + echo begin + eval set -- $(false) + echo end Because of this, almost every bug report mentions something along the lines of getopt printed an error. This is getout outputting to stderr, - but the exit code is being swallowed and errexit is not happening. + but the non-zero exit code is being swallowed and errexit is not + happening. This coding was probably designed intentionally to disable automatic exit whenever getopt found an unexpected argument, due to the fact that this (currently) happens on almost every kickstart file as our syntax support is quite limited. However, it also hides the warning from the user, which I believe is the real bug. Moreover, the only error is printed briefly to a terminal and then never seen again, which makes it tough for people to put in bug reports. Perhaps a better solution would be to use set +x and set -e to temporarily disable errexit, and use a trap on ERR to ensure that these errors are caught and printed to both syslog and the current TTY. -- You received this bug notification because you are a member of Ubuntu Bugs, which is subscribed to Ubuntu. https://bugs.launchpad.net/bugs/1331814 Title: Warning messages for handlers are never called To manage notifications about this bug go to: https://bugs.launchpad.net/ubuntu/+source/kickseed/+bug/1331814/+subscriptions -- ubuntu-bugs mailing list ubuntu-bugs@lists.ubuntu.com https://lists.ubuntu.com/mailman/listinfo/ubuntu-bugs
[Bug 776701] Re: firewall command not implemented
** Changed in: kickseed (Ubuntu) Status: Triaged = Confirmed -- You received this bug notification because you are a member of Ubuntu Bugs, which is subscribed to Ubuntu. https://bugs.launchpad.net/bugs/776701 Title: firewall command not implemented To manage notifications about this bug go to: https://bugs.launchpad.net/ubuntu/+source/kickseed/+bug/776701/+subscriptions -- ubuntu-bugs mailing list ubuntu-bugs@lists.ubuntu.com https://lists.ubuntu.com/mailman/listinfo/ubuntu-bugs
[Bug 1330799] [NEW] No --log option with %post sections
Public bug reported: The %post section in RHEL has a --log option that kickseed does not support. Oddly enough, there should be a generic warning printed about this - kickseed/handlers/post.sh should pick up that --log is unrecognized and call warn_getopt %post. I don't see that warning being printed - perhaps it's not showing up in /var/log/installer/syslog for some reason? Just for future testing, here are some cases that don't work: %post --nochroot --log=/target/test1.log echo 'Hello, World' %end %post --nochroot --log=/test2.log echo 'Hello, World' %end %post --log=/target/test3.log echo 'Hello, World' %end %post --log=/test4.log echo 'Hello, World' %end %post exec /dev/tty3 /dev/tty3 chvt 3 echo echo echo # Running Post Configuration # echo ( echo 'Hello, World!' ) 21 | /usr/bin/tee /var/log/test5.log chvt 1 %end Also, the following does not work. I suspect I'm making some mistake here, as this should work just fine with the current kickseed codebase if I understand correctly. However this switching TTY stuff is new to me, so perhaps I'm mistaking something %post --nochroot exec /dev/tty3 /dev/tty3 chvt 3 echo echo echo # Running Post Configuration # echo ( echo 'Hello, World!' ) 21 | /usr/bin/tee /target/test6.log chvt 1 %end ** Affects: kickseed (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/1330799 Title: No --log option with %post sections To manage notifications about this bug go to: https://bugs.launchpad.net/ubuntu/+source/kickseed/+bug/1330799/+subscriptions -- ubuntu-bugs mailing list ubuntu-bugs@lists.ubuntu.com https://lists.ubuntu.com/mailman/listinfo/ubuntu-bugs