Matahari August 2011

matahari@lists.fedorahosted.org

22 participants
72 discussions

Whitespace
by Andrew Beekhof 24 Aug '11

24 Aug '11

You'll be happy to know I have enabled the following from http://www.jwz.org/doc/tabs-vs-spaces.html for all C/C++ files [quote] To cause TAB characters to not be used in the file for compression, and for only spaces to be used, do this: (setq indent-tabs-mode nil) [/quote] And I still only have to press the TAB-key so everyone can be happy.

1 0

Documentation for Matahari users
by Russell Bryant 24 Aug '11

24 Aug '11

Greetings, Documentation is absolutely critical for the success of our project. I've spent some time thinking about documentation for our users. Here is my proposal for the direction we should take. ---- Existing Documentation First, it's worth looking at what we have so far. We have 2 documentation sets up and running so far: a) http://matahariproject.org/dev-manual/ b) http://clusterlabs.org/doxygen/matahari/1f5dfea/index.html There is not necessarily overlap between the content, but the audience for these two sets of documentation is the same: people working on matahari itself. If all goes as planned, there will be a much larger audience that needs documentation from us: those that write applications against the agents we provide. ---- To-Do The gap in the documentation we have today is for the users of the matahari agents. I don't think it's in the scope of our project to document how to write applications using QMF or DBus in general. However, we can help fill in the gaps in this area as we see fit. One of the biggest things we can do here is provide good examples in our project. It is definitely within the scope of our project to make it quick and easy to figure out the details of the APIs we provide. Andrew has done some nice work demonstrating how an XSLT can be used to transform the schema files into another format. Take a look: $ xsltproc doc/api.xsl src/host/schema.xml I think we can take this, run with it, and end up with some great docs. Proposed steps: 1) Continue adding documentation in the schema. For some items, the schema already provides a place to include a description. For more verbose descriptions of methods, etc., we can place it in an XML comment above the entry in the schema. The host schema has some examples of this and the existing XSLT demonstrates how we can grab the contents of the comment. We should go through and add documentation for _everything_ in the schemas. 2) Use the existing docbook manual (dev-manual in links above) as the base for the documentation. Move the existing content to be a "Writing a new agent" chapter. 3) Create two new chapters: - Chapter 1: Project Overview Go over the project from a user's point of view. List the agents that we have and at a high level, what they provide. - Matahari APIs The APIs we provide. The core of this chapter can be automatically generated using an XSLT similar to doc/api.xsl. Instead of transforming into a plain text format, it could generate docbook XML. Publican does a great job of rendering stellar looking documentation from docbook. If we complete these steps, I think we'll have a great manual that addresses our entire audience. As always, feedback welcome. Thanks, -- Russell Bryant

3 4

matahari added you to a team
by GitHub 24 Aug '11

24 Aug '11

Commit grants 8 users push access to 3 repositories. View team: https://github.com/organizations/matahari/teams/48010 -- Reply to this email directly or view it on GitHub: https://github.com/organizations/matahari/teams/48010

1 0

[IANA #483358] Application for port-number: matahari
by Pearl Liang via RT 24 Aug '11

24 Aug '11

Thank you for the reply. > Assignee Name (REQUIRED): Matahari Project > Assignee E-mail (REQUIRED): matahari(a)lists.fedorahosted.org Please confirm if http://matahariproject.org/ is the organization's website. On Wed Aug 24 06:07:57 2011, zbitter(a)redhat.com wrote: > Assignee Name (REQUIRED): Matahari Project > Assignee E-mail (REQUIRED): matahari(a)lists.fedorahosted.org > (The Assignee is the organization, company or individual person > responsible > for the initial assignment. ) > Contact Name (REQUIRED): Zane Bitter > Contact E-mail (REQUIRED): zbitter(a)redhat.com > > Transport Protocol(s) (REQUIRED): (Please indicate which protocols are > currently required.) > [ X ] TCP [ ] UDP [ ] SCTP [ ] DCCP > > Service Code (REQUIRED for DCCP only) > > A: ____________________________ > > Service Name (REQUIRED, 15 character maximum): > > A: matahari > > Desired Port Number (OPTIONAL, for Port Number only) > > A: 49000 > NOTE: It is *inappropriate* to use a port number until your > application has > been approved for assignment. Leave it blank if no number preference. > > Description (REQUIRED) (For example: Super User Messaging) > > A: Matahari Broker > > Reference (REQUIRED): Please provide a brief and basic technical > description > of the protocol that will use the service name or port number, > including > message > formats, types, sequences, functionalities, of your protocol. In > addition, > please address the specific questions included below to the extent > possible. > > A: The Matahari server is an implementation of an Advanced Message > Queuing Protocol (AMQP) broker. Matahari uses a subset of the Qpid > Management Framework (QMF), which sits on top of AMQP (Apache Qpid is > an > implementation of AMQP). AMQP is an open standard which is documented > here: > http://www.amqp.org/confluence/download/attachments/720900/amqp.0- > 10.pdf > > In short, AMQP is a generic protocol for handling message queues. > Client > applications can create and subscribe to queues and publish messages > to > them. > > Matahari (http://www.matahariproject.org/) is a service to provide > remote access to system management APIs over QMF. Existing APIs > support > system configuration and querying of system parameters for virtual > servers, however the protocol is extensible so users can add their own > custom APIs. > > Q: If broadcast/multicast is used, how and what for? > > A: Multicast may optionally be used to efficiently combine multiple > brokers into a high-availability cluster. > > Q: If UDP is requested, please explain how traffic is limited, and > whether the protocol reacts to congestion. > > A: > ___________________________________________________________________________ > > Q: If UDP is requested, please indicate whether the service is solely > for the discovery of hosts supporting this protocol. > > A: > ___________________________________________________________________________ > > Q: Please explain how your protocol supports versioning. > > A: All AMQP messages begin with a version field, and AMQP defines a > method for the client and server to negotiate a mutually compatible > protocol version. QMF objects have a schema which can be queried over > the protocol. > > Q: If your request is for more than one transport, please explain in > detail how the protocol differs over each transport. > > A: > __________________________________________________________________________ > > Q: Please describe how your protocol supports security. Please note > that > there is current debate as to whether additional future ports will be > allocated for newly assigned services that lack security, e.g., solely > for the addition of a secure variant. Applicants are strongly advised > to > address this issue as early as possible in the design of their > protocols. > > > A: The protocol may optionally be tunnelled over SSL. > > > Q: Please explain the state of development of your protocol. > > A: AMQP is a published open standard with four published versions > since > 2006 and multiple available implementations. QMF has undergone a major > revision and is now considered stable. > > Q: If SCTP is requested, please describe the following: Is there an > existing TCP and/or UDP service name or port number assignment? > If so, please indicate the details below: > > Service Name: _______________; Port number: _______________ > > What specific SCTP capability is used by the application such that a > user who has the choice of both TCP (and/or UDP) and SCTP ports for > this application would choose SCTP? See [RFC4960 section 7.1] > (http://tools.ietf.org/html/rfc4960#section-7.1) > > A: > _____________________________________________________________________________ > > ________________________________________________________________________________ > > > Please provide any other information that would be helpful in > understanding how this protocol differs from existing assigned > services. > > A: Matahari differs from a generic AMQP broker in that it requires the > QMF protocol running on top, so only the Apache Qpid implementation of > AMQP can be used. Since Matahari is designed for system configuration, > for security reasons it places a tighter than usual set of > restrictions > on how the connection can be used, and thus must be segregated from > regular applications' AMQP traffic. >

2 1

[PATCH v2] Fix compiler warnings in Windows builds
by Zane Bitter 24 Aug '11

24 Aug '11

Signed-off-by: Zane Bitter <zbitter(a)redhat.com> --- src/lib/host_windows.c | 25 +++++++++----------- src/lib/network_windows.c | 56 ++++++++++++++------------------------------ src/lib/services_windows.c | 6 ++--- 3 files changed, 33 insertions(+), 54 deletions(-) diff --git a/src/lib/host_windows.c b/src/lib/host_windows.c index 0a599d3..149cc2a 100644 --- a/src/lib/host_windows.c +++ b/src/lib/host_windows.c @@ -44,27 +44,29 @@ host_os_get_cpu_flags(void) } static void -get_token_priv(HANDLE token, TOKEN_PRIVILEGES tkp) +enable_se_priv(void) { - OpenProcessToken(GetCurrentProcess(), - TOKEN_ADJUST_PRIVILEGES | TOKEN_QUERY, &token); + HANDLE token; + TOKEN_PRIVILEGES tkp; + LookupPrivilegeValue(NULL, SE_SHUTDOWN_NAME, &tkp.Privileges[0].Luid); tkp.PrivilegeCount = 1; tkp.Privileges[0].Attributes = SE_PRIVILEGE_ENABLED; - AdjustTokenPrivileges(token, FALSE, &tkp, 0, - (PTOKEN_PRIVILEGES) NULL, 0); + if (OpenProcessToken(GetCurrentProcess(), + TOKEN_ADJUST_PRIVILEGES | TOKEN_QUERY, &token)) { + AdjustTokenPrivileges(token, FALSE, &tkp, 0, + (PTOKEN_PRIVILEGES) NULL, 0); + CloseHandle(token); + } } void host_os_reboot(void) { - HANDLE token; - TOKEN_PRIVILEGES tkp; - - get_token_priv(token, tkp); + enable_se_priv(); ExitWindowsEx(EWX_REBOOT | EWX_FORCE, SHTDN_REASON_FLAG_PLANNED); } @@ -72,10 +74,7 @@ host_os_reboot(void) void host_os_shutdown(void) { - HANDLE token; - TOKEN_PRIVILEGES tkp; - - get_token_priv(token, tkp); + enable_se_priv(); ExitWindowsEx(EWX_SHUTDOWN | EWX_FORCE, SHTDN_REASON_FLAG_PLANNED); } diff --git a/src/lib/network_windows.c b/src/lib/network_windows.c index 8e017ad..fe300e7 100644 --- a/src/lib/network_windows.c +++ b/src/lib/network_windows.c @@ -25,64 +25,44 @@ #endif #include "matahari/network.h" +#include "matahari/utilities.h" #include <config.h> #include <glib.h> #include <windows.h> -void -network_os_stop(const char *iface) +static void +network_os_setstate(const char *iface, const char *state) { PROCESS_INFORMATION pi; STARTUPINFO si; char *p; char *exe_path; + wchar_t *wexe_path; p = getenv("WINDIR"); - exe_path = g_strdup_printf("%s\\system32\\netsh interface set interface " - "%s disabled", p, iface); + exe_path = g_strdup_printf("\"%s\\system32\\netsh\" " + "interface set interface %s %s", + p, iface, state); + wexe_path = char2wide(exe_path); + g_free(exe_path); - gboolean ok = CreateProcess(NULL, - exe_path, - NULL, - NULL, - TRUE, - 0, - NULL, - NULL, - &si, - &pi); + CreateProcess(NULL, wexe_path, NULL, NULL, TRUE, 0, NULL, NULL, &si, &pi); CloseHandle(pi.hThread); CloseHandle(pi.hProcess); - g_free(exe_path); + free(wexe_path); } void -network_os_start(const char *iface) +network_os_stop(const char *iface) { - PROCESS_INFORMATION pi; - STARTUPINFO si; - char *p; - char *exe_path; - - p = getenv("WINDIR"); - exe_path = g_strdup_printf("%s\\system32\\netsh interface set interface " - "%s enabled", p, iface); - - gboolean ok = CreateProcess(NULL, - exe_path, - NULL, - NULL, - TRUE, - 0, - NULL, - NULL, - &si, - &pi); + network_os_setstate(iface, "disabled"); +} - CloseHandle(pi.hThread); - CloseHandle(pi.hProcess); - g_free(exe_path); +void +network_os_start(const char *iface) +{ + network_os_setstate(iface, "enabled"); } diff --git a/src/lib/services_windows.c b/src/lib/services_windows.c index 80b9a39..b48805d 100644 --- a/src/lib/services_windows.c +++ b/src/lib/services_windows.c @@ -109,10 +109,9 @@ static int extract_service_status(const char *unmutable, int max) { int lpc = 0, last = 0; - int rc = LSB_STATUS_OTHER_ERROR; if (unmutable == NULL || max < 10) { - return rc; + return LSB_STATUS_OTHER_ERROR; } for (; lpc < max; lpc++) { @@ -147,6 +146,7 @@ extract_service_status(const char *unmutable, int max) rc = LSB_STATUS_OK; break; default: + rc = LSB_STATUS_OTHER_ERROR; mh_info("Unknown status: %s", next); } /* free(next); */ @@ -160,7 +160,7 @@ extract_service_status(const char *unmutable, int max) } } mh_info("Couldn't parse service status"); - return rc; + return LSB_STATUS_OTHER_ERROR; } gboolean

2 1

[PATCH] Fix compiler warnings in Windows builds
by Zane Bitter 24 Aug '11

24 Aug '11

Signed-off-by: Zane Bitter <zbitter(a)redhat.com> --- src/lib/host_windows.c | 25 +++++++++----------- src/lib/network_windows.c | 56 ++++++++++++++------------------------------ src/lib/services_windows.c | 5 ++-- 3 files changed, 32 insertions(+), 54 deletions(-) diff --git a/src/lib/host_windows.c b/src/lib/host_windows.c index 0a599d3..149cc2a 100644 --- a/src/lib/host_windows.c +++ b/src/lib/host_windows.c @@ -44,27 +44,29 @@ host_os_get_cpu_flags(void) } static void -get_token_priv(HANDLE token, TOKEN_PRIVILEGES tkp) +enable_se_priv(void) { - OpenProcessToken(GetCurrentProcess(), - TOKEN_ADJUST_PRIVILEGES | TOKEN_QUERY, &token); + HANDLE token; + TOKEN_PRIVILEGES tkp; + LookupPrivilegeValue(NULL, SE_SHUTDOWN_NAME, &tkp.Privileges[0].Luid); tkp.PrivilegeCount = 1; tkp.Privileges[0].Attributes = SE_PRIVILEGE_ENABLED; - AdjustTokenPrivileges(token, FALSE, &tkp, 0, - (PTOKEN_PRIVILEGES) NULL, 0); + if (OpenProcessToken(GetCurrentProcess(), + TOKEN_ADJUST_PRIVILEGES | TOKEN_QUERY, &token)) { + AdjustTokenPrivileges(token, FALSE, &tkp, 0, + (PTOKEN_PRIVILEGES) NULL, 0); + CloseHandle(token); + } } void host_os_reboot(void) { - HANDLE token; - TOKEN_PRIVILEGES tkp; - - get_token_priv(token, tkp); + enable_se_priv(); ExitWindowsEx(EWX_REBOOT | EWX_FORCE, SHTDN_REASON_FLAG_PLANNED); } @@ -72,10 +74,7 @@ host_os_reboot(void) void host_os_shutdown(void) { - HANDLE token; - TOKEN_PRIVILEGES tkp; - - get_token_priv(token, tkp); + enable_se_priv(); ExitWindowsEx(EWX_SHUTDOWN | EWX_FORCE, SHTDN_REASON_FLAG_PLANNED); } diff --git a/src/lib/network_windows.c b/src/lib/network_windows.c index 8e017ad..fe300e7 100644 --- a/src/lib/network_windows.c +++ b/src/lib/network_windows.c @@ -25,64 +25,44 @@ #endif #include "matahari/network.h" +#include "matahari/utilities.h" #include <config.h> #include <glib.h> #include <windows.h> -void -network_os_stop(const char *iface) +static void +network_os_setstate(const char *iface, const char *state) { PROCESS_INFORMATION pi; STARTUPINFO si; char *p; char *exe_path; + wchar_t *wexe_path; p = getenv("WINDIR"); - exe_path = g_strdup_printf("%s\\system32\\netsh interface set interface " - "%s disabled", p, iface); + exe_path = g_strdup_printf("\"%s\\system32\\netsh\" " + "interface set interface %s %s", + p, iface, state); + wexe_path = char2wide(exe_path); + g_free(exe_path); - gboolean ok = CreateProcess(NULL, - exe_path, - NULL, - NULL, - TRUE, - 0, - NULL, - NULL, - &si, - &pi); + CreateProcess(NULL, wexe_path, NULL, NULL, TRUE, 0, NULL, NULL, &si, &pi); CloseHandle(pi.hThread); CloseHandle(pi.hProcess); - g_free(exe_path); + free(wexe_path); } void -network_os_start(const char *iface) +network_os_stop(const char *iface) { - PROCESS_INFORMATION pi; - STARTUPINFO si; - char *p; - char *exe_path; - - p = getenv("WINDIR"); - exe_path = g_strdup_printf("%s\\system32\\netsh interface set interface " - "%s enabled", p, iface); - - gboolean ok = CreateProcess(NULL, - exe_path, - NULL, - NULL, - TRUE, - 0, - NULL, - NULL, - &si, - &pi); + network_os_setstate(iface, "disabled"); +} - CloseHandle(pi.hThread); - CloseHandle(pi.hProcess); - g_free(exe_path); +void +network_os_start(const char *iface) +{ + network_os_setstate(iface, "enabled"); } diff --git a/src/lib/services_windows.c b/src/lib/services_windows.c index 80b9a39..8344d2d 100644 --- a/src/lib/services_windows.c +++ b/src/lib/services_windows.c @@ -109,10 +109,9 @@ static int extract_service_status(const char *unmutable, int max) { int lpc = 0, last = 0; - int rc = LSB_STATUS_OTHER_ERROR; if (unmutable == NULL || max < 10) { - return rc; + return LSB_STATUS_OTHER_ERROR; } for (; lpc < max; lpc++) { @@ -160,7 +159,7 @@ extract_service_status(const char *unmutable, int max) } } mh_info("Couldn't parse service status"); - return rc; + return LSB_STATUS_OTHER_ERROR; } gboolean

1 0

Re: [Matahari] IANA Port assignment application
by Russell Bryant 24 Aug '11

24 Aug '11

Moved to the ML. On 08/23/2011 03:10 PM, Perry Myers wrote: > On 08/22/2011 01:07 PM, Zane Bitter wrote: >> Folks, >> We need to get the IANA to assign matahari a TCP port so that we can get >> into /etc/services. The process for doing so is documented here: >> http://tools.ietf.org/html/draft-ietf-tsvwg-iana-ports-10 > > excellent, I'm glad to see this getting pushed forward. Though... if > we're going to get a different port than 49000 we probably need to know > sooner rather than later since switching ports once things are 'out > there' and in use will be difficult, no? There's not a great way to handle it. At some point we need to make a release that uses the new port number by default. Any older versions that want to connect to newer stuff will have to use the --port options. Release documentation can help that a bit. The sooner we do it the better, for sure. -- Russell Bryant

2 1

Updated Backlog page
by Russell Bryant 23 Aug '11

23 Aug '11

Greetings, The Backlog page has been updated to make it a bit easier to see what is still on the to-do list, what is currently in progress, and what has already been completed. Take a look here: https://github.com/matahari/matahari/wiki/Backlog Thanks, -- Russell Bryant

1 0

Patch workflow
by Russell Bryant 23 Aug '11

23 Aug '11

Greetings, I would like to offer up an alternative patch workflow for discussion. Currently we're using the mailing list for posting patches and reviews. How would you feel about using a code review tool to manage the patch workflow? The main reason I have for proposing the change is to help ensure that things don't get dropped and that it's very clear to everyone what the status is of each pending patch. Specifically, the tool I have in mind is Review Board [1]. fedorahosted has a reviewboard instance up that (I presume) we could use [2]. Q&A: 1) Well what about keeping discussions on the mailing list? If we used this tool, we could still have code review discussions echoed to the mailing list so everyone still sees what's going on. As an example, the qpid project is using Review Board [3] and review requests are sent to their dev mailing list [4]. 2) How do you post patches? Here is a general workflow overview [5]. Posting patches is pretty easy to do from the command line. There is a tool called post-review [6] that you use. 3) I hate the web. If I can't do it from my terminal, it's terrible. Ok, well that's not a question. I do respect your right to have an opinion, even if it's wrong. I do believe that there are some tools coming in future versions of Review Board that will allow you to do everything from the command line, but it's not there yet. [1] http://www.reviewboard.org/ [2] http://fedorahosted.org/reviewboard/ [3] https://reviews.apache.org/groups/qpid/ [4] http://mail-archives.apache.org/mod_mbox/qpid-dev/201108.mbox/browser [5] http://www.reviewboard.org/docs/manual/1.5/users/getting-started/workflow/ [6] http://www.reviewboard.org/docs/manual/1.5/users/tools/post-review/#post-re… Let me know what you think! Thanks, -- Russell Bryant

4 6

[PATCH] Resolve errors reported by coverity.
by Russell Bryant 23 Aug '11

23 Aug '11

[PATCH 1/4] core: rework dead code. [PATCH 2/4] core: Remove unused variable. This is not a coverity reported error. I just pulled out the formatting cleanup from the real code changes in patch 4/4. [PATCH 3/4] host: Clean up indentation. [PATCH 4/4] qmf: Don't use string after it has been destroyed. -- Russell Bryant

1 4

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

2013

2012

2011

2010

Matahari August 2011