cockpit-devel November 2017

cockpit-devel@lists.fedorahosted.org

8 participants
10 discussions

Host Devices
by Marek Libra 27 Jan '18

27 Jan '18

Hi, I would like to renew the effort on Host Devices for Cockpit [1]. To move forward, let's implement it step by step while opening brand new PR picking just a subset of the already implemented functionality [1] while adjusted to generally acceptable form. Initial implementation would meet: - PCI support only - initially read-only: just the List of devices by their Class (according to [2]) - example: by Audio device, Ethernet Controller, etc - data source: sysfs - make use of backward-compatible lspci for data preprocessing (especially manipulation with HW database) - monitor for changes by listening kernel uevents Follow-ups will lead to [3] scope: - for pci, active actions are allowed for selected device classes, one PR per each: - (un)bind VFIO driver - configure SR-IOV network cards - vGPU configuration - additional view by Driver - What devices is the driver bound to? - additional view by IOMMU Groups - For server fine-tuning, what devices are within single IOMMU Group? - support for other buses, i.e. USB or SCSI - their views are independent on each other due to fundamental differences Please let me know your thoughts. Thanks, Marek [1] https://github.com/cockpit-project/cockpit/pull/5523 [2] http://pci-ids.ucw.cz/ [3] https://github.com/cockpit-project/cockpit/wiki/Feature:-Hardware-Devices -- Marek Libra senior software engineer Red Hat Czech <https://www.redhat.com>

4 8

Does cockpit offer a scp feature?
by Mark Reynolds 30 Nov '17

30 Nov '17

Hello, Is there a cockpit API/feature that allows you to copy files from one machine to another (like scp)? I didn't seen anything in the docs. I could do a cockpit.spawn and use scp that way(I think), but I'm wondering if there is a better way. Thanks, Mark

2 1

Manage local SELinux policy in Cockpit
by Petr Lautrbach 21 Nov '17

21 Nov '17

Hi, there's already an epic in trello related to Improved SELinux troubleshooting and Management [1]. There seems to be missing a user story for SELinux management so I put together a document which should cover basic SELinux local policy management using either semanage command or org.selinux DBUS interface: https://plautrba.fedorapeople.org/manage-local-selinux-policy-in-cockpit.ht… I'd like to ask you for a review and comments if it makes sense and for help with design for this effort when there's an agreement [1] https://trello.com/c/WiFrlt4C/381-epic-improved-selinux-troubleshooting-and… Thanks, Petr

4 9

Re: Why would cockpit.spawn() cause my js function to disappear? Resolved
by Mark Reynolds 17 Nov '17

17 Nov '17

On 11/17/2017 01:35 PM, Peter wrote: > Add some console.log lines to your done and fail functions and before > your "// Set html paragragh to insts". It might help you get a better > idea of what is happening. > > with the .done call you are providing a callback that gets called when > the spawn command finishes. At that point you now have the result and > you can do what you want with it. > > So if your goal here is to run that command and then write it to html > that needs to happen inside the done callback. Uggggh, of course. It was working earlier because I was actually updating a HTML from the "done" function, but then I tried to move that result output to a drop down list. However, I was doing it from outside of the done callback function. Thank you very much for clarifying the behavior, and immediately identifying my misunderstanding of how the done function worked. Mark > > cockpit.spawn(['ls', '/tmp'], { superuser: true}) > .done(function(data) { > // use data to update html > }).fail(function(error){ > // write a failure message to html > }); > > > On 11/17/2017 10:24 AM, Mark Reynolds wrote: >> >> >> On 11/17/2017 12:50 PM, Peter wrote: >>> cockpit.spawn returns a promise. It's asynchronous. get_insts will >>> return before the callbacks that change the value of insts get run. >>> You need to rethink what you trying to do here and base your code off >>> of events or callbacks. >> But I thought the done() method was supposed to wait? Okay... So please >> tell me how do I get cockpit to run a os command, and get the >> results(stdout) from that command? >> >> What is weird is that cockpit.spawn().done() is working in other JS >> functions I have, except this one. Very odd... Still trying to see >> what I changed that broke things because it was working fine a few hours >> ago... >> >>> >>> >>> On 11/17/2017 09:40 AM, Mark Reynolds wrote: >>>> Okay I'm really stumped because this was just working, but >>>> cockpit.spawn >>>> seems to be misbehaving >>>> >>>> Here is my js: >>>> >>>> function get_insts() { >>>> var insts = ["None"]; >>>> >>>> cockpit.spawn(['ls', '/tmp'], { superuser: true >>>> }).done(function(data) { >>>> insts= ["okay we actually got here"]; >>>> }).fail(function(error){ >>>> insts = ["FAIL"]; >>>> }); >>>> // Set html paragragh to insts >>>> ... >>>> } >>>> >>>> $(window.document).ready(function() { >>>> get_insts(); >>>> $("body").show(); >>>> }); >>>> >>>> >>>> >>>> "insts" is always set to "None". Debugging shows odd behavior. >>>> When I >>>> refresh the page cockpit.spawn just doesn't do anything. >>>> >>>> cockpit.js >>>> >>>> 1832 function(e ,n) <= n is the correct list of arguments >>>> >>>> But it does not execute function, then it skips to >>>> >>>> 1842 u.close = function(e) { >>>> >>>> It does not execute function, then it skips to >>>> >>>> 1847 returns u >>>> >>>> We don't fail, and we don't succeed, and my js function completes. >>>> Then... my js function get_insts() gets called again automatically >>>> (not >>>> sure how its getting called twice from the page reload), and it does >>>> appear to succeed the second time, data has expected values, but >>>> somehow >>>> after we get the data the debugger stops in a cockpit function listed >>>> below. It loops a few times in the cockpit function, then my js >>>> function, get_insts(), just aborts and disappears silently even though >>>> things seemed to working correctly on the second pass. >>>> >>>> cockpit-ws-151-2.fc26 >>>> >>>> cockpit.js:867 >>>> >>>> function P(e) { >>>> var n, t, r; >>>> r = e.pending; >>>> e.process_scheduled = false; >>>> e.pending = undefined; >>>> for (var o = 0, a = r.length; o < a; ++o) { >>>> e.pur = true; >>>> t = r[o][0]; >>>> n = r[o][e.status]; >>>> if (i(n)) { >>>> t.resolve(n.apply(e.promise, e.values)); >>>> <-- >>>> debugger stops here, then its loops and moves on >>>> } else if (e.status === 1) { >>>> t.resolve.apply(t.resolve, e.values); >>>> } else { >>>> t.reject.apply(t.reject, e.values); >>>> } >>>> } >>>> } >>>> >>>> >>>> I don't know what to do next, this is very odd behavior. Especially >>>> since it worked earlier today. I even rebooted my laptop :-( I tried >>>> different and simple commands to send to cockpit.spawn, but it doesn't >>>> change this behavior. >>>> >>>> Anything else I should be looking at? >>>> >>>> Thanks, >>>> Mark >>>> _______________________________________________ >>>> cockpit-devel mailing list -- cockpit-devel(a)lists.fedorahosted.org >>>> To unsubscribe send an email to >>>> cockpit-devel-leave(a)lists.fedorahosted.org >>>> >>> >>> >>> >> >

1 0

Why would cockpit.spawn() cause my js function to disappear?
by Mark Reynolds 17 Nov '17

17 Nov '17

Okay I'm really stumped because this was just working, but cockpit.spawn seems to be misbehaving Here is my js: function get_insts() { var insts = ["None"]; cockpit.spawn(['ls', '/tmp'], { superuser: true }).done(function(data) { insts= ["okay we actually got here"]; }).fail(function(error){ insts = ["FAIL"]; }); // Set html paragragh to insts ... } $(window.document).ready(function() { get_insts(); $("body").show(); }); "insts" is always set to "None". Debugging shows odd behavior. When I refresh the page cockpit.spawn just doesn't do anything. cockpit.js 1832 function(e ,n) <= n is the correct list of arguments But it does not execute function, then it skips to 1842 u.close = function(e) { It does not execute function, then it skips to 1847 returns u We don't fail, and we don't succeed, and my js function completes. Then... my js function get_insts() gets called again automatically (not sure how its getting called twice from the page reload), and it does appear to succeed the second time, data has expected values, but somehow after we get the data the debugger stops in a cockpit function listed below. It loops a few times in the cockpit function, then my js function, get_insts(), just aborts and disappears silently even though things seemed to working correctly on the second pass. cockpit-ws-151-2.fc26 cockpit.js:867 function P(e) { var n, t, r; r = e.pending; e.process_scheduled = false; e.pending = undefined; for (var o = 0, a = r.length; o < a; ++o) { e.pur = true; t = r[o][0]; n = r[o][e.status]; if (i(n)) { t.resolve(n.apply(e.promise, e.values)); <-- debugger stops here, then its loops and moves on } else if (e.status === 1) { t.resolve.apply(t.resolve, e.values); } else { t.reject.apply(t.reject, e.values); } } } I don't know what to do next, this is very odd behavior. Especially since it worked earlier today. I even rebooted my laptop :-( I tried different and simple commands to send to cockpit.spawn, but it doesn't change this behavior. Anything else I should be looking at? Thanks, Mark

1 0

Cockpit 156 released
by Martin Pitt 16 Nov '17

16 Nov '17

http://cockpit-project.org/blog/cockpit-156.html Cockpit is the modern Linux admin interface. We release regularly. Here are the release notes from version 156. Redesign main navigation and support mobile browsing ---------------------------------------------------- The top-level dashboard menus and second-level menus of them have been redesigned to clarify the structure, be more extensible, and use less vertical screen space. This version introduces proper support for using on mobile phones and tablets. As part of this, the login screen was rearranged in mobile mode to avoid overlapping the on-screen keyboard with the user/password input fields: Screenshots: http://cockpit-project.org/images/mobilenav-desktop.png http://cockpit-project.org/images/mobilenav-mobile-menu.png http://cockpit-project.org/images/mobilenav-login.png Add project homepage links to Apps page --------------------------------------- Applications on the Apps page now show a link to the project home page, if provided in the AppStream metadata: Thanks Benjamin Deering for this feature! Screenshot: http://cockpit-project.org/images/apps-homepage-link.png Support alternate Kerberos keytabs ---------------------------------- If present, Cockpit will now use the Kerberos `/etc/cockpit/krb5.keytab` file instead of the default system keytab `/etc/krb5.keytab`. See the Single Sign On documentation for details. Maintain an /etc/issue(5) file with current Cockpit status ---------------------------------------------------------- Cockpit now maintains a `/run/cockpit/issue` file with instructions how to enable Cockpit (when disabled) or the Cockpit URL (when enabled). The next util-linux 2.32 release will support [1] `/etc/issue.d/*.issue` drop-ins. Distributions or users can make use of this by installing an `/etc/issue.d/01-cockpit.issue` symlink to the above file. [1] https://github.com/karelzak/util-linux/commit/1fc82a1360305f696dc1be6105c9c… Use event-driven refresh of oVirt virtual machine data instead of polling ------------------------------------------------------------------------- This makes the page much more reactive and use less resources. Thanks to Marek Libra for this improvement! Get it ------ You can get Cockpit here: http://cockpit-project.org/running.html Cockpit 156 is available in Fedora 27: https://bodhi.fedoraproject.org/updates/cockpit-156-1.fc27 Or download the tarball here: https://github.com/cockpit-project/cockpit/releases/tag/156 Take care, Martin Pitt

1 0

Odd cockpit.spawn() behavior
by Mark Reynolds 15 Nov '17

15 Nov '17

Basically if I try to use wildcards or pipes things seem to break: var cmd = ["/usr/bin/ls", "/etc/dirsrv/slapd-*"]; cockpit.spawn(cmd, { superuser: true }).done(function(data) { insts = new String(data); $("#Title").text(data); } ).fail(function(error){ $("#Title").text(error); }); I get this error response: "/usr/bin/ls exited with code 2" I get the same error when trying this: var cmd = ["/usr/bin/ls", "/etc/dirsrv/", "|", "/usr/bin/grep", "slapd"]; But this works fine: var cmd = ["/usr/bin/ls", "/etc/dirsrv/"]; Is this a bug, or do I need to do something different? Thanks, Mark

2 2

cannot download diagnostic report
by Fuji San 13 Nov '17

13 Nov '17

After I have done running the diagnostic report, the blue "download report" button shows up. But when I click on it, a popup windows asks me if I want to save the file"sos-report-hostname-20171026143147.tar.xz". When I click "Save", the popup closes and well nothing else happens! No file is saved. The browser does not show any downloaded file.

2 3

Need help with new package: JS vs CSP
by Mark Reynolds 10 Nov '17

10 Nov '17

Hello, I'm evaluating cockpit to be a potential UI replacement for our database server (389 directory server). I am also not an experienced web developer, but regardless I'm going to be writing our new UI. I've created a sample package(manifest.json file, html files, etc) and I put it all under ~/.local/share/cockpit/test/ I see my new menu item, and it loads my html page. But I can not get my custom javascript working on my project page: Content Security Policy: The page’s settings blocked the loading of a resource at self (“default-src http://localhost:9090 http://localhost:9090”). Source: onclick attribute on BUTTON element. I tried putting my javascript in a separate file, and embedding it in the html page, etc. None it works and it's the CSP that "appears" to be blocking me every time. What is the correct way to get my javascript functions working in my package? Thanks, Mark

2 4

Cockpit 155 release
by Martin Pitt 08 Nov '17

08 Nov '17

http://cockpit-project.org/blog/cockpit-155.html Cockpit is the modern Linux admin interface. We release regularly. Here are the release notes from version 155. NFS client support ------------------ The Storage page now shows currently mounted NFS volumes and offers to mount/unmount them. Note that this does not yet support Kerberos authentication; for now this relies on the classic `hosts.allow`/`hosts.deny` ACLs. Video: https://www.youtube.com/watch?v=PGCBda3Le9Y Add "Maintenance" switch for oVirt hosts ---------------------------------------- Switching a host to maintenance mode causes all VMs to be migrated to another running oVirt host. Screnshot: http://cockpit-project.org/images/ovirt-maintenance.png Fix Terminal rendering issues in Chrome --------------------------------------- Console programs like Midnight Commander which use graphical characters were previously distorted in the Terminal in the Chrome Browser. Improve Ctrl+W behaviour in the Terminal ---------------------------------------- Ctrl+W is commonly being used in a Terminal shell to delete the last word, but in web browsers it closes the current window or tab. Cockpit now causes a confirmation dialog to pop up on Ctrl+W when the Terminal is focussed to avoid accidental closing. Support the upcoming OpenShift 3.7 release ------------------------------------------ When available, use the `rbac.authorization` API instead of the `policybinding` API for determining project permissions. OpenShift 3.7 now properly supports the former and stops supporting the latter. Cockpit remains compatible with older OpenShift releases. Get it ------ You can get Cockpit here: http://cockpit-project.org/running.html Cockpit 155 is available in Fedora 27: https://bodhi.fedoraproject.org/updates/cockpit-155-1.fc27 Or download the tarball here: https://github.com/cockpit-project/cockpit/releases/tag/155 Take care, Martin Pitt

1 0

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

2013

cockpit-devel November 2017