Re: OFW access from linux
deepak wrote: On Sep 15 2009, at 17:31, Paul Fox was caught saying: john wrote: there's no SysRq key on the XO keyboard, so you'll need to use a break on the serial console to invoke it... Please. If you're going to put this hook in, which I think is a great idea, at least make it work on the standard hardware! And when the operating system is not very responsive. That's when you'll need it for debugging or resetting the system. Without taking the plastic apart, finding a part with no known suppliers, and soldering it to your motherboard. i agree. getting the hook in place was a first step, and satisfies my immediate need for wanting to be able to look at and modify h/w registers during development -- the /proc trigger is sufficient for that. we have a constant tension between just make it work on the XO and make it so we have a prayer of merging it upstream without lots of pushback, and i confess i erred towards the latter, this time. custom key combinations would require more extensive changes to the input drivers than i was willing to take on last night. ... There are some pretty obscure keys on the XO keyboard -- howabout something like holding down the leftmost and rightmost gradually increasing sized dots keys in the top row, and then pressing the M-for-Mitch key? i would applaud, encourage, and otherwise buy beer for anyone that could help out in making something like this happen. (i might get to it someday, myself, but that will no doubt be _after_ i need it.) I disagree. sysrq is the standard way of trapping keyboard sequences to cause special kernel actions. If you want another set but the issue is that there's no such key on the XO keyboard: SysRq/PrtSc doesn't exist. of key sequences to make this happen, you probably want to trap it in X and have that send the proper sequnce to the /proc. I guess don't see the benefit of moving from a simple two key combination to some multi-key command. it did occur to me earlier that adding a hook such as you describe to olpc-kbdshim (which is watching all mouse and keyboard traffic anyway) might be useful. if the system is hung, it won't help, but it might in many cases. In the case that the kernel is hung to the point that we can't talk to it at all, if the hang is caused by an oops, we can have a patch that traps calls to die() and bug() and jumps to OFW. This is what KGDB does, so there is some precedent. We could also have some magic in the EC to look for a sequence of keys that signals a low-level handler to kick into OFW in the case of major lockup (assuming there's enough space in EC code space to do this...). interesting idea: an EC solution would be the most failsafe, but i don't believe it currently actually looks at any of the keystroke data as it passes through. (except for the keys on the screen bezel, which it has complete responsibility for.) i think we'd have to see a pretty strong need to go down this path, though. I've not seen many total hangs that can't be fixed with some printk debugging + KGDB when it works. Or a JTAG unit, but I don't think we can do that on the XO, right? right, no jtag. but as you say, cases where the system is well and truly hung and need sysrq aren't all that frequent. paul ~Deepak =- paul fox, p...@laptop.org ___ Devel mailing list Devel@lists.laptop.org http://lists.laptop.org/listinfo/devel
Re: XO-1 Mesh support in F11, testing request
2009/9/16 Daniel Drake d...@laptop.org: all rpms at http://dev.laptop.org/~dsd/20090916/ found a few minutes to find a few problems: forgot to install the fdi file olpcmesh.py is not being installed from the sugar package shortened dhcpcd timeout not working Sugar never seems to give up trying school mesh (or never falls back to link local) no time to investigate further now Daniel ___ Devel mailing list Devel@lists.laptop.org http://lists.laptop.org/listinfo/devel
Re: XO-1 Mesh support in F11, testing request
Haven't had a chance to try this yet, been busy with assessment tasks ... but what particular F11 for XO-1 build are you testing with? -- James Cameron http://quozl.linux.org.au/ ___ Devel mailing list Devel@lists.laptop.org http://lists.laptop.org/listinfo/devel
B2 build status
When Quanta finally started the build at 7PM on Wed. they discovered two problems: A - zcat to one of the three candidate microSD cards were taking four to six times longer than it should have. B - The units were powering off occasionally when turning on the WLAN card. They had to stop production, and rework all the motherboards to solve B (I had proposed the solution used on Monday, but they tried a different one first). This was completed this afternoon, and 120 reworked machines were put into a twelve hour burn-in (with many WLAN card power cycles) over Thursday night. Mitch, cjb, pgf, dsd, Quozl, smithbone, and I worked on problem A all day, and haven't found the root cause yet. As we are reasonably sure that there should be a software solution, Quanta is switching to use fs-update in their manufacturing process and will ship these machines to us for further testing.The SD card in question is one that they had mistakenly not provided to me ahead of time for qualification. Quanta will restart the production line Friday morning to build the remainder of the B2 laptops, and they should be shipped to 1CC on Sunday. Salutations from Song Jiang, wad ___ Devel mailing list Devel@lists.laptop.org http://lists.laptop.org/listinfo/devel
lease not fount
Como estas martin he tengo un problema te lo explico en el colegio que entregamos en el fidel coloma realizamos una visita anti-ayer por lo que dos niños se hacercaron a mi con sus xo. cuando ensiendo la XO manda la pantalla de Invalid System Date y mas habajo lease expired y empieza hacer el procedimiento de busqueda en los dispositivo de almacenamineto ej:(MicroSD, USB y Wifi) entonces lo que yo hago es saco el serial y el uuid de la XO y hago un archivo csv con los datos antes mecionado y despues me voy hacerle el nuevo leasig.sig en la maquina firmadora, una ves hecho eso y generado el leasig.sig, ahora introdusco la usb en la XO que esta bloqueada y la coloca a la hora que empieza a buscar los disitivo y lo acepta luego quito la usb y queda lista, pero lo mas raro de todo es que cuando la reinicio me vuelve aparecer el mesaje que antes me mostraba de Invalid System Date + lease expired. que pasa que no me esta copiando la leasig. -- Kevin Mauricio Benavides Castro Fundación Zamora Terán www.fundacionzt.org Technical Support-Nicaragua OLPC - www.fedoraproject.org Fedora Project Developer Linux User: #12356 ___ Devel mailing list Devel@lists.laptop.org http://lists.laptop.org/listinfo/devel
Re: lease not fount
2009/9/17 Kevin Mauricio Benavides Castro webmaster.ac@gmail.com: cuando ensiendo la XO manda la pantalla de Invalid System Date y mas habajo lease expired y empieza hacer el procedimiento de busqueda en los dispositivo de almacenamineto ej:(MicroSD, USB y Wifi) entonces lo que yo hago es saco el serial y el uuid de la XO y hago un archivo csv con los datos antes mecionado y despues me voy hacerle el nuevo leasig.sig en la maquina firmadora, una ves hecho eso y generado el leasig.sig, ahora Ok. Tienes el procedimiento casi correcto. Para corregir la fecha cuando el XO dice 'invalid system date' necesitas un devkey. En vez de generar el lease, tienes que generar un 'devkey' (hay un programa identico al que genera leases, que genera un devkey para esa maquina). Con el devkey en un USB, arrancas la maquina y desde open firmware corriges el reloj -- instrucciones aqui: http://wiki.laptop.org/go/Fix_Clock#Developer_Key_Procedure_-_for_symptom_2_.28screen_turns_on.29 (No hace falta poner un firmware nuevo, el que tienen esas maquinas ya es Q2E41) Ahora _esta no es la solución completa_. Esas máquinas estaban andando bien la semana anterior cuando ustedes las prepararon, lo cual significa que la batería del reloj está mala, o está un poco suelta. Mañana esa XO va a tener le mismo problema. Así que lo que tienes que hacer es llevar - un destornillador - el pegamento especial que tienen para arreglar el soporte para la bateria - una bateria de reemplazo -- la puedes comprar en ferreterías, lleva una bateria de muestra. - un cepillo de dientes Cuando abres la XO, lo mas importante es ver que la bateria no haya perdido el jugo corrosivo. - Si hay corrosión http://wiki.laptop.org/go/Fix_Clock#Fixing_the_leaking_battery_problem - Si la bateria esta 'floja' en el soporte http://wiki.laptop.org/go/Fix_Clock#Fixing_the_loose_battery_problem Puede que no tenga corrosión y que el soporte esté bien, y que simplemente la batería esté vieja y haya que cambiarla. abrazos, m -- martin.langh...@gmail.com mar...@laptop.org -- School Server Architect - ask interesting questions - don't get distracted with shiny stuff - working code first - http://wiki.laptop.org/go/User:Martinlanghoff ___ Devel mailing list Devel@lists.laptop.org http://lists.laptop.org/listinfo/devel
Re: OFW access from linux
On Thu, Sep 17, 2009 at 05:10:06AM -0400, Paul Fox wrote: but the issue is that there's no such key on the XO keyboard: SysRq/PrtSc doesn't exist. Key to the left of frame and above erase is still free isn't it? ;-) The legend matches my idea of switching to OFW. Another system hiding in the background. Still, kernel handling a keyboard key still requires some sanity on the part of the kernel. -- James Cameron http://quozl.linux.org.au/ ___ Devel mailing list Devel@lists.laptop.org http://lists.laptop.org/listinfo/devel
Re: [Server-devel] Attempting to re-install jabber component of XS without re-installing the entire server
Martin, I ran : yum --enablerepo=olpcxs-testing update It downloaded overnight and I ran the upgrade this morning. It both seemed to work and seemed to fail. Failure: during the install of the kernal module there was an I/O error and I believe a second failure cat /etc/issue returns: OLPC School Server release 9-0.5.2 Kernel \r on an \m (\l) is this what I should see after update? Success: The install as a whole seems to have completed successfully. # uname -a returns Linux schoolserver.lccnjimeta.org 2.6.27.25-78.2.56.fc9.i686 #1 SMP TH Jun 18 12:47:50 EDT 2009 i686 i686 i386 GNU/Linux I am guessing that the fc9.686 means fedora 9? That would be success right? when I try to re-run the yum --enablerepo... again it returns: Setting up Upddate Process No Packages marked for Update After running the update ejabberd was still crashing when I started it. However, I hadn't followed the instructions for clearing out ejabberd yet from your first reply. I tried following those steps (after upgrading to .6) and they seem to have worked... The service runs now without crashing, I can ping schoolserver.lccnjimeta.org and 172.18.0.1 successfully #hostname -f returns schoolserver.lccnjimeta.org However, even though I can ping these from my laptop, I can't access the web based jabber interface. I'm certain that they are communicating further, because when I open http://172.18.01/ (without :5280/admin) it leads me to a page telling me that moodle is currently disabled. Any ideas? Thanks again! -Dan On Wed, Sep 16, 2009 at 2:56 PM, Martin Langhoff martin.langh...@gmail.comwrote: On Wed, Sep 16, 2009 at 3:36 PM, Daniel Bennett dant...@gmail.com wrote: I installed what I think was 0.5.2 which was the newest stable release available when I brought it to Jimeta in late March of 09. What does the following command say? cat /etc/issue So the steps that I should take are: 1) backup some part of my server (how should I go about doing this?) Are we talking about making a copy of a particular directory here, or running a utility of some kind? Uh, so many ways to skin this cat! You will need an external disk for this (usb?) and probably a LiveUSB disk (or LiveCD). If you are familiar with cloning Linux OSs with rsync, that's one path (that has a quota of DIY). If you want an easy-to-use answer, I hear Mondo Rescue is popular and good, but I haven't used it myself. Your 2,3,4 steps are right... 5) Configure my magically working Jabbery goodness Step 5 is actually Enjoy magic workingness And how stable is this .6 pre-release candidate, because I want this to be as stable as possible when I leave. Definitely more stable than 0.5.2 which messed up your ejabberd install :-) m -- martin.langh...@gmail.com mar...@laptop.org -- School Server Architect - ask interesting questions - don't get distracted with shiny stuff - working code first - http://wiki.laptop.org/go/User:Martinlanghoff ___ Server-devel mailing list Server-devel@lists.laptop.org http://lists.laptop.org/listinfo/server-devel
Re: [Server-devel] Attempting to re-install jabber component of XS without re-installing the entire server
I've replied in line below. On Thu, Sep 17, 2009 at 11:12 AM, Martin Langhoff martin.langh...@gmail.com wrote: On Thu, Sep 17, 2009 at 11:51 AM, Daniel Bennett dant...@gmail.com wrote: I ran : yum --enablerepo=olpcxs-testing update It downloaded overnight and I ran the upgrade this morning. Good... It both seemed to work and seemed to fail. Failure: during the install of the kernal module there was an I/O error and I believe a second failure Hmmm. An I/O error sounds a serious problem. You can verify a successful install of the kernel by doing: rpm -V kernel it should return without printing anything (no news: good news). dan: I got no news, so I guess it worked cat /etc/issue returns: OLPC School Server release 9-0.5.2 Kernel \r on an \m (\l) is this what I should see after update? that's normal until I formally release it :-) Success: The install as a whole seems to have completed successfully. good! After running the update ejabberd was still crashing when I started it. That is odd. dan: you may have answered this before reading the rest of my email, but if it was unclear, ejabberd is no longer crashing The service runs now without crashing, I can ping schoolserver.lccnjimeta.org and 172.18.0.1 successfully #hostname -f returns schoolserver.lccnjimeta.org However, even though I can ping these from my laptop, I can't access the web based jabber interface. The webbased interface is disabled now. No longer needed. dan: I ran: ejabberdctl connected-users It returns long output like (the random string here is made up by me): 34209381afaa07fa9df...@schoolserver.lccnjimeta.org/Telepathy and the number of users has grown some this morning, from about 4 to over 10 (we have a number of XO's in use here this morning for continued teacher training) So I definitely buy that some automagic is going on here. Is there perhaps documentation I can view that explains to me how to create different groups for different classes and generally manage things without the now obsolete web interface? (real world example: I would like to put all of my 5th graders in a group together, my 2nd graders in another, etc) I might even want to change these on a regular basis. Is this all done through moodle perhaps? Is there documentation I should be using for that? If you could send me a link it would be much appreciated. I'm certain that they are communicating further, because when I open http://172.18.01/ (without :5280/admin) it leads me to a page telling me that moodle is currently disabled. Moodle is disabled? Did you restart the machine? Can you tell us the output of the commands below? service pgsql-xs status Dan: postmaster is stopped chkconfig --list pgsql-xs Dan: service pgsql-xs supports chkconfig, but is not reerenced in any runlevel (run 'chkconfig --add pgsql-xs') Should I run this as it says? Other oddities noticed during reboot if they help..: klogctl during boot returns 'invalid argument' mshbonds and wmeshs do not seem to be present could not connect to server postgres as well. -D m -- martin.langh...@gmail.com mar...@laptop.org -- School Server Architect - ask interesting questions - don't get distracted with shiny stuff - working code first - http://wiki.laptop.org/go/User:Martinlanghoff ___ Server-devel mailing list Server-devel@lists.laptop.org http://lists.laptop.org/listinfo/server-devel
Re: [Server-devel] timestamps on backups
Thanks Tom and Martin, I think I have figured it out. Sorry for any confusion. It seems the timestamps are correct on the server, it is only when displayed in the backup page (dsbackup.php) where the times are off. For example, a backup done just a few minutes ago would say it was done 4 hours ago on that page. The problem is the timezone. When I specifically set the date.timezone parameter in php.ini to UTC, the text started appearing correctly in the backup page. So it seems PHP is using a different timezone. /etc/sysconfig/clock says that the timezone of the system is America/New York which is probably the timezone PHP is using. On Wed, 2009-09-16 at 15:05 -0700, Tom Mitchell wrote: On Wed, Sep 16, 2009 at 5:37 AM, Martin Langhoff martin.langh...@gmail.com wrote: On Wed, Sep 16, 2009 at 2:28 PM, Hamilton Chua hamilton.c...@gmail.com wrote: I'm not sure if anybody has noticed this yet but after doing a backup, it seems the datestamps on the backup page are wrong. We might need more detail than that if we're to understand the situation :-) Is there something about time zones and time synchronization that we need to be aware of with regards to backups. Yes, the utc time on all machines should make sense. So do date --utc on XS and on the laptops involved to make sure all players are in the same decade. This has been tested with machines on different TZs so it should work. As long as utc agrees across machines. -- martin.langh...@gmail.com http://lists.laptop.org/listinfo/server-devel The two numbers 4 hours and 40 years are almost telling. Linux keeps time as seconds from midnight January 1, 1970 12:00:00 GMT. Today GMT (Greenwich Mean Time) has been replaced by UTS (Coordinated Universal Time) which is GMT done better with atomic clocks. Since you are using XS on 0.6d5 and Sugar on a Stick the system gets the initial time of day (date) from the local hardware clock. A unix/Linux system default sets the local hardware clock to UTS while windows sets it to local time. Depending on daylight savings time in (say) Oklahoma four hours looks like a Windows system setting the hardware time of day. Since Linux can be configured to play nice with windows and set the hardware clock to local time windows is not always the issue in possiblly confusing the offset from GMT/UTS. The 40 years is very close to the beginnig of unix time (zero seconds) and can be seen on a confused local time of day clock. NTP (network time protocol) tools can be used to set the time of day on a network connected system to the correct UTS time. Local time is computed based on UTS and an offset time zone. Since all binary time stamps are UTS different users can set different timezone values in their environment and the system will do the 'right' thing. See: date date --uts date -u (export TZ=Europe/Paris;date;date-u) # touch /tmp/now stat /tmp/now (export TZ=Europe/Paris; stat now) ___ Server-devel mailing list Server-devel@lists.laptop.org http://lists.laptop.org/listinfo/server-devel
Re: [Server-devel] timestamps on backups
On Thu, Sep 17, 2009 at 1:13 PM, Hamilton Chua hamilton.c...@gmail.com wrote: /etc/sysconfig/clock says that the timezone of the system is America/New York which is probably the timezone PHP is using. That's weird. Questions - Can you confirm date --utc is correct on the XS and on the involved XOs? - If you look at the datastore-timestamp directories - are the timestamps correctly in UTC? - If you run tzselect, change the tz to your local tz -- does it fix the problem? If it does, is the problem fixed for new backups or for all backups? The PHP code in dsbackup.php should be ignoring timezones, but something might be sneaking in. Also - the timestamp in the dirnames may be wrong. m -- martin.langh...@gmail.com mar...@laptop.org -- School Server Architect - ask interesting questions - don't get distracted with shiny stuff - working code first - http://wiki.laptop.org/go/User:Martinlanghoff ___ Server-devel mailing list Server-devel@lists.laptop.org http://lists.laptop.org/listinfo/server-devel
Re: [Server-devel] Troubles running F9 mock chroot under F11
2009/9/16 Jerry Vonau jvo...@shaw.ca: That should be do-able using mkslim (read it first) from xs-livecd's git repo, along with my idea to use a pre-configured updates repo on the iso. http://lists.laptop.org/pipermail/server-devel/2009-February/002937.html Thanks! Got it working as follows: 1. extract ISO 2. copy in new ks file 3. add more RPMs to Packages/ (using creative use of yumdownloader to make sure that deps come with the new RPMs) 4. createrepo --database --groupfile repodata/comps.xml . 5. remove stuff that mkslim removes 6. mkisofs no need to mess with bdb stuff any more :) Daniel ___ Server-devel mailing list Server-devel@lists.laptop.org http://lists.laptop.org/listinfo/server-devel
Re: [Server-devel] Attempting to re-install jabber component of XS without re-installing the entire server
On Thu, Sep 17, 2009 at 12:37 PM, Daniel Bennett dant...@gmail.com wrote: rpm -V kernel dan: I got no news, so I guess it worked Good! dan: I ran: ejabberdctl connected-users It returns long output like (the random string here is made up by me): 34209381afaa07fa9df...@schoolserver.lccnjimeta.org/Telepathy That all sounds right. Some more useful commands # will show 'Online' exists $ejabberdctl srg-list-groups `hostname -f` Online # show the settings of the Online Shared Roster Group - $ ejabberdctl srg-get-info 'Online' `hostname -f` name: Online displayed_groups: [Online] description: Created_by_ejabberd_init online_users: true members: So I definitely buy that some automagic is going on here. Hey -- did you ever doubt it? :-) Is there perhaps documentation I can view that explains to me how to create different groups for different classes and generally manage things without the now obsolete web interface? (real world example: I would like to put all of my 5th graders in a group together, my 2nd graders in another, etc) I might even want to change these on a regular basis. Is this all done through moodle perhaps? Is there documentation I should be using for that? If you could send me a link it would be much appreciated. Yes, you can segregate what XOs see in the neighbourhood view by moodle course membership. This is mainly useful when you have schools with more than ~50 XOs. Quick howto: 1 - Make sure you are using a recent Browse.xo (101 at least, I think) 2 - The first XO to register (reboot) and visit moodle successfully gets some extra rights (is an 'admin' of sorts). So get an XO reg'd and visiting Moodle -- it will auto-authenticate into Moodle. Should see a 'site administration' block on the left. 3 - Create some courses (Site Administration-courses-Add/edit-Add new), assign teachers and students. In the long and confusing new course form, all you need to set is Full Name and Short Name. Please ignore every other option. Note: when it comes to enrolling, setup some courses for XOs you have registered already -- this is mainly to have something to work with! You can add more courses enrolments later. Might help: http://docs.moodle.org/en/Enrolment#Manual_enrolment 4 - In 'Site Administration', go to Courses-Presence Service, and set presencebycourse to Yes. 5 - Within 10 minutes, issuing the same ejabberdctl queries as above should show that Online has been replaced by several SRGs -- one per course. Asking for the 'info' of those, will show you their membership. 6 - When you make the switch from one mode to the other on the XS side, the situation will be confusing for the XOs, so they might need to re-associate to the Access Point (so they re-query their group membership) before they see the changes. Note: All changes to the course membership take 5~10 minutes to appear in ejabberd on the XOs, and some changes may also need an XO to reassociate to the AP. Emphasis on 'some' :-) Dan: service pgsql-xs supports chkconfig, but is not reerenced in any runlevel (run 'chkconfig --add pgsql-xs') Should I run this as it says? Yes! and then restart the server. Both pgsql-xs and moodle will start up. On this first startup, moodle will perform its setup (so it'll take a tad longer). You can see the log of the install in /var/log/moodle/instupg.log klogctl during boot returns 'invalid argument' dunno mshbonds and wmeshs do not seem to be present normal. could not connect to server postgres as well. not normal but you'll have it fixed for the next reboot :-) m -- martin.langh...@gmail.com mar...@laptop.org -- School Server Architect - ask interesting questions - don't get distracted with shiny stuff - working code first - http://wiki.laptop.org/go/User:Martinlanghoff ___ Server-devel mailing list Server-devel@lists.laptop.org http://lists.laptop.org/listinfo/server-devel
Re: [Server-devel] timestamps on backups
Hi Martin, My responses inline ... - Can you confirm date --utc is correct on the XS and on the involved XOs? Yes, date--utc on both XS and XO's are the same if not close to each other, the difference being a few seconds. - If you look at the datastore-timestamp directories - are the timestamps correctly in UTC? Yes, the directory name uses the UTC timestamp. - If you run tzselect, change the tz to your local tz -- does it fix the problem? If it does, is the problem fixed for new backups or for all backups? The timezone of the system has always been the local tz. The PHP code in dsbackup.php should be ignoring timezones, but something might be sneaking in. Also - the timestamp in the dirnames may be wrong. I checked the directory names and it seems they are correctly UTC. It appears that the problem is just the calculation involved that returns how long ago the backup was. It seems the time() function returns the epoch based on the local timezone instead of the epoch based on UTC. Best, Ham ___ Server-devel mailing list Server-devel@lists.laptop.org http://lists.laptop.org/listinfo/server-devel
Re: [Server-devel] timestamps on backups
On Thu, Sep 17, 2009 at 1:55 PM, Hamilton Chua hamilton.c...@gmail.com wrote: - Can you confirm date --utc is correct on the XS and on the involved XOs? Yes, date--utc on both XS and XO's are the same if not close to each other, the difference being a few seconds. thanks I checked the directory names and it seems they are correctly UTC. Thanks! It seems the time() function returns the epoch based on the local timezone instead of the epoch based on UTC. Actually - time() is doiing the right thing. It was the use of mktime() that was bogus. Replace mktime() with gmmtime(), the 3 times it appears in dsbackuplib.php and the problem is gone. thanks for the debugging! m -- martin.langh...@gmail.com mar...@laptop.org -- School Server Architect - ask interesting questions - don't get distracted with shiny stuff - working code first - http://wiki.laptop.org/go/User:Martinlanghoff ___ Server-devel mailing list Server-devel@lists.laptop.org http://lists.laptop.org/listinfo/server-devel
Re: [Server-devel] timestamps on backups
Actually - time() is doiing the right thing. It was the use of mktime() that was bogus. Replace mktime() with gmmtime(), the 3 times it appears in dsbackuplib.php and the problem is gone. Thanks Martin :-) thanks for the debugging! You're welcome. ___ Server-devel mailing list Server-devel@lists.laptop.org http://lists.laptop.org/listinfo/server-devel
Re: [Server-devel] Troubles running F9 mock chroot under F11
On Thu, Sep 17, 2009 at 1:47 PM, Daniel Drake d...@laptop.org wrote: Thanks! Got it working as follows: 1. extract ISO 2. copy in new ks file 3. add more RPMs to Packages/ (using creative use of yumdownloader to make sure that deps come with the new RPMs) 4. createrepo --database --groupfile repodata/comps.xml . 5. remove stuff that mkslim removes 6. mkisofs no need to mess with bdb stuff any more :) Good to hear it's worked! Thinking about future support for that deployment and also for the XS (as upstream) -- is it in your plans to document what rpms and ks changes you are using? cheers, m -- martin.langh...@gmail.com mar...@laptop.org -- School Server Architect - ask interesting questions - don't get distracted with shiny stuff - working code first - http://wiki.laptop.org/go/User:Martinlanghoff ___ Server-devel mailing list Server-devel@lists.laptop.org http://lists.laptop.org/listinfo/server-devel
Re: [Server-devel] Attempting to re-install jabber component of XS without re-installing the entire server
We saw this message at boot, not sure if it's relevant: AICCU is not configured go to /etc/aiccu pssql now starts on boot, but we are still getting the same results when accessing: http://172.18.0.1/(moodle/) The result is a page stating: Moodle is disabled at the moment. We have rebooted more than once at this point. We've been digging around (but not changing) /etc/httpd/conf/httpd.conf and /etc/httpd/conf/httpd-xs.conf and they both point to a document root of /var/www/html Does this have anything to do with our problem? I don't really know what I'm doing... So, in a nutshell, pssql is running at startup, but we still can't access moodle through the web interface. Any ideas? Thanks yet again, -Dan (and Michael Ishaku) On Thu, Sep 17, 2009 at 12:54 PM, Martin Langhoff martin.langh...@gmail.com wrote: On Thu, Sep 17, 2009 at 12:37 PM, Daniel Bennett dant...@gmail.com wrote: rpm -V kernel dan: I got no news, so I guess it worked Good! dan: I ran: ejabberdctl connected-users It returns long output like (the random string here is made up by me): 34209381afaa07fa9df...@schoolserver.lccnjimeta.org/Telepathy That all sounds right. Some more useful commands # will show 'Online' exists $ejabberdctl srg-list-groups `hostname -f` Online # show the settings of the Online Shared Roster Group - $ ejabberdctl srg-get-info 'Online' `hostname -f` name: Online displayed_groups: [Online] description: Created_by_ejabberd_init online_users: true members: So I definitely buy that some automagic is going on here. Hey -- did you ever doubt it? :-) Is there perhaps documentation I can view that explains to me how to create different groups for different classes and generally manage things without the now obsolete web interface? (real world example: I would like to put all of my 5th graders in a group together, my 2nd graders in another, etc) I might even want to change these on a regular basis. Is this all done through moodle perhaps? Is there documentation I should be using for that? If you could send me a link it would be much appreciated. Yes, you can segregate what XOs see in the neighbourhood view by moodle course membership. This is mainly useful when you have schools with more than ~50 XOs. Quick howto: 1 - Make sure you are using a recent Browse.xo (101 at least, I think) 2 - The first XO to register (reboot) and visit moodle successfully gets some extra rights (is an 'admin' of sorts). So get an XO reg'd and visiting Moodle -- it will auto-authenticate into Moodle. Should see a 'site administration' block on the left. 3 - Create some courses (Site Administration-courses-Add/edit-Add new), assign teachers and students. In the long and confusing new course form, all you need to set is Full Name and Short Name. Please ignore every other option. Note: when it comes to enrolling, setup some courses for XOs you have registered already -- this is mainly to have something to work with! You can add more courses enrolments later. Might help: http://docs.moodle.org/en/Enrolment#Manual_enrolment 4 - In 'Site Administration', go to Courses-Presence Service, and set presencebycourse to Yes. 5 - Within 10 minutes, issuing the same ejabberdctl queries as above should show that Online has been replaced by several SRGs -- one per course. Asking for the 'info' of those, will show you their membership. 6 - When you make the switch from one mode to the other on the XS side, the situation will be confusing for the XOs, so they might need to re-associate to the Access Point (so they re-query their group membership) before they see the changes. Note: All changes to the course membership take 5~10 minutes to appear in ejabberd on the XOs, and some changes may also need an XO to reassociate to the AP. Emphasis on 'some' :-) Dan: service pgsql-xs supports chkconfig, but is not reerenced in any runlevel (run 'chkconfig --add pgsql-xs') Should I run this as it says? Yes! and then restart the server. Both pgsql-xs and moodle will start up. On this first startup, moodle will perform its setup (so it'll take a tad longer). You can see the log of the install in /var/log/moodle/instupg.log klogctl during boot returns 'invalid argument' dunno mshbonds and wmeshs do not seem to be present normal. could not connect to server postgres as well. not normal but you'll have it fixed for the next reboot :-) m -- martin.langh...@gmail.com mar...@laptop.org -- School Server Architect - ask interesting questions - don't get distracted with shiny stuff - working code first - http://wiki.laptop.org/go/User:Martinlanghoff ___ Server-devel mailing list Server-devel@lists.laptop.org http://lists.laptop.org/listinfo/server-devel
Re: [Server-devel] Attempting to re-install jabber component of XS without re-installing the entire server
On Thu, Sep 17, 2009 at 4:11 PM, Daniel Bennett dant...@gmail.com wrote: We saw this message at boot, not sure if it's relevant: AICCU is not configured go to /etc/aiccu That's normal. pssql now starts on boot, but we are still getting the same results when accessing: http://172.18.0.1/(moodle/) Just to confirm, what is the output of...? service pgsql-xs status (we expect it to be running) service moodle status (Moodle cron job is enabled) chkconfig --list pgsql-xs (should say that it's on in 3,4 and 5) chkconfig --list moodle (same) Note! There is a 'postgres' service too. That should be stopped. The result is a page stating: Moodle is disabled at the moment. That is because Moodle is having trouble starting up -- and that is all about getting the 'pgsql-xs' service going, and the initial moodle setup running. Some more debugging: - Does the file /etc/moodle/needsupgrade exist? (It should :-) ) - Does the file /var/log/moodle-instupg.log exist? You'll want to look at it. Lines with 'PHP Notice' are safe to ignore, you should see lots of SQL, and a final line saying [Wed Sep 16 05:12:57 EDT 2009] Finished install / upgrade - Success. That file probably has the secret to your woes... - What happens if you do...? /etc/initd.d/moodle stop (bash -x /etc/initd.d/moodle start 21) | tee moodleinit.log The 'tee' command will show you the output but will also save it to the moodleinit.log file. So I am hoping you can post your /var/log/moodle-instupg.log and moodleinit.log files so we can debug further. We've been digging around (but not changing) /etc/httpd/conf/httpd.conf and /etc/httpd/conf/httpd-xs.conf That should not affect things. If moodle is showing you the Moodle is disabled at the moment msg, the apache config is doing its job just fine :-) and they both point to a document root of /var/www/html we are overriding it -- see the /etc/httpd/conf.d/ directory :-) Does this have anything to do with our problem? I don't really know what I'm doing... You're doing pretty well, I gotta say - m -- martin.langh...@gmail.com mar...@laptop.org -- School Server Architect - ask interesting questions - don't get distracted with shiny stuff - working code first - http://wiki.laptop.org/go/User:Martinlanghoff ___ Server-devel mailing list Server-devel@lists.laptop.org http://lists.laptop.org/listinfo/server-devel
Re: [Server-devel] Attempting to re-install jabber component of XS without re-installing the entire server
On Thu, Sep 17, 2009 at 3:30 PM, Martin Langhoff martin.langh...@gmail.comwrote: On Thu, Sep 17, 2009 at 4:11 PM, Daniel Bennett dant...@gmail.com wrote: We saw this message at boot, not sure if it's relevant: AICCU is not configured go to /etc/aiccu That's normal. pssql now starts on boot, but we are still getting the same results when accessing: http://172.18.0.1/(moodle/) http://172.18.0.1/%28moodle/%29 Just to confirm, what is the output of...? service pgsql-xs status (we expect it to be running) fine service moodle status (Moodle cron job is enabled) DISABLED chkconfig --list pgsql-xs (should say that it's on in 3,4 and 5) fine chkconfig --list moodle (same) fine Note! There is a 'postgres' service too. That should be stopped. couldn't find this service The result is a page stating: Moodle is disabled at the moment. That is because Moodle is having trouble starting up -- and that is all about getting the 'pgsql-xs' service going, and the initial moodle setup running. Some more debugging: - Does the file /etc/moodle/needsupgrade exist? (It should :-) ) exists - Does the file /var/log/moodle-instupg.log exist? exists You'll want to look at it. Lines with 'PHP Notice' are safe to ignore, you should see lots of SQL, and a final line saying [Wed Sep 16 05:12:57 EDT 2009] Finished install / upgrade - Success. FAILURE: mentions of correcting version/local_version mangling - will error out on fresh DBs UPDATE 0 EnvironmentPHP Fatal error: Call to undefined function upgrade_fix_incorrect_mnethostides() in /var/moodle/web/lib/db/upgrade.php on line 3167 [Time date] Finished install / upgrade - Failure I haven't answered any questions below this line. We are trying to find a way to get these files off the server. I can't figure out how to set it up as an FTP server, I can't get any linux email sending to work ((al)pine, and mutt don't seem to be installed, and when I use mail I get a message telling me that root doesn't have mail) We are trying to set another machine up to SSH into it (a machine with email access) but no luck. Any idea how I can send an email from the server? Thanks, -Dan That file probably has the secret to your woes... - What happens if you do...? /etc/initd.d/moodle stop (bash -x /etc/initd.d/moodle start 21) | tee moodleinit.log The 'tee' command will show you the output but will also save it to the moodleinit.log file. So I am hoping you can post your /var/log/moodle-instupg.log and moodleinit.log files so we can debug further. We've been digging around (but not changing) /etc/httpd/conf/httpd.conf and /etc/httpd/conf/httpd-xs.conf That should not affect things. If moodle is showing you the Moodle is disabled at the moment msg, the apache config is doing its job just fine :-) and they both point to a document root of /var/www/html we are overriding it -- see the /etc/httpd/conf.d/ directory :-) Does this have anything to do with our problem? I don't really know what I'm doing... You're doing pretty well, I gotta say - m -- martin.langh...@gmail.com mar...@laptop.org -- School Server Architect - ask interesting questions - don't get distracted with shiny stuff - working code first - http://wiki.laptop.org/go/User:Martinlanghoff ___ Server-devel mailing list Server-devel@lists.laptop.org http://lists.laptop.org/listinfo/server-devel
Re: [Server-devel] Attempting to re-install jabber component of XS without re-installing the entire server
http://sprunge.us/NcQT http://sprunge.us/MERg That curl to sprunge.us trick is nifty. automatically posting text files to a web page on the net via shell / command line .. I should have asked that question a long time ago :) On Thu, Sep 17, 2009 at 4:48 PM, Martin Langhoff martin.langh...@gmail.comwrote: On Thu, Sep 17, 2009 at 5:30 PM, Daniel Bennett dant...@gmail.com wrote: EnvironmentPHP Fatal error: Call to undefined function upgrade_fix_incorrect_mnethostides() in /var/moodle/web/lib/db/upgrade.php on line 3167 [Time date] Finished install / upgrade - Failure That might be the culprit. Very weird. Still, want to see the logs... We are trying to set another machine up to SSH into it (a machine with email access) but no luck. Any idea how I can send an email from the server? email is a bit awkward, but there is a pastebin-style service we can use. Try cat /path/to/file | curl -F 'sprunge=-' http://sprunge.us After a few seconds it will reply with a URL, like http://sprunge.us/EJJF where the file can be found. /etc/initd.d/moodle stop (bash -x /etc/initd.d/moodle start 21) | tee moodleinit.log I really want that moodleinit.log :-) cheers, m -- martin.langh...@gmail.com mar...@laptop.org -- School Server Architect - ask interesting questions - don't get distracted with shiny stuff - working code first - http://wiki.laptop.org/go/User:Martinlanghoff ___ Server-devel mailing list Server-devel@lists.laptop.org http://lists.laptop.org/listinfo/server-devel
Re: [Server-devel] Attempting to re-install jabber component of XS without re-installing the entire server
On Thu, Sep 17, 2009 at 6:23 PM, Daniel Bennett dant...@gmail.com wrote: http://sprunge.us/NcQT http://sprunge.us/MERg Hmmm. Very odd! Can you check that your moodle-xs package is correct? What's the output of rpm -q moodle-xs rpm -V moodle-xs (should be empty) That curl to sprunge.us trick is nifty. Sure is good. I googled it up 5 minutes ago :-) m -- martin.langh...@gmail.com mar...@laptop.org -- School Server Architect - ask interesting questions - don't get distracted with shiny stuff - working code first - http://wiki.laptop.org/go/User:Martinlanghoff ___ Server-devel mailing list Server-devel@lists.laptop.org http://lists.laptop.org/listinfo/server-devel
Re: [Server-devel] Attempting to re-install jabber component of XS without re-installing the entire server
On Thu, Sep 17, 2009 at 7:11 PM, Daniel Bennett dant...@gmail.com wrote: rpm -q moodle-xs moodle-xs-1.9.5.xs1.1.gc292d55-1.xs9.noarch rpm -V moodle-xs (should be empty) returns nothing Thanks. I had to re-read your logs a bit. In brief: I made the mistake of reading the moodle-instupg.log bottom-up, and thought the install had become corrupted. But re-visiting it, I realise that the error that is key here is the failure to connect to the Postgres server. psql: could not connect to server: No such file or directory Is the server running locally and accepting connections on Unix domain socket /tmp/.s.PGSQL.5432? So we need to make sure that the pgsql-xs service is running, and that it accepts connections from Moodle (that is, from apache). Some commands to try (and the correct output in my test XS) # check that the postgresql service is *not* set to run as it will conflict w pgsql-xs $ chkconfig --list postgresql postgresql 0:off 1:off 2:off 3:off 4:off 5:off 6:off # should say that the server is running $ service pgsql-xs status # should give you a list of databases - see how moodle-xs is owned by apache $ su postgres -c '/usr/bin/psql -l ' List of databases Name| Owner | Encoding +--+--- moodle-xs | apache | UTF8 postgres | postgres | SQL_ASCII template0 | postgres | SQL_ASCII template1 | postgres | SQL_ASCII # check that you can connect as the apache user and list existing tables $sudo -u apache psql moodle-xs -c '\d' List of relations Schema | Name | Type | Owner +-+--+ public | adodb_logsql| table| apache public | adodb_logsql_id_seq | sequence | apache public | mdl_assignment | table| apache public | mdl_assignment_id_seq | sequence | apache public | mdl_assignment_submissions | table| apache public | mdl_assignment_submissions_id_seq | sequence | apache public | mdl_backup_config | table| apache ... Can you try these commands? I think we're pretty close to resolving the problem. m -- martin.langh...@gmail.com mar...@laptop.org -- School Server Architect - ask interesting questions - don't get distracted with shiny stuff - working code first - http://wiki.laptop.org/go/User:Martinlanghoff ___ Server-devel mailing list Server-devel@lists.laptop.org http://lists.laptop.org/listinfo/server-devel
[Server-devel] DNS/dhcp question
Hello. I have the XS server up and running on a Dell. Per specs, I believe, I have two NICs, which I can see as eth0 and eth1. Eth0 is connected to the internet. Eth1 is connected to the AP for the XOs. Or, more accurate, the AP for the XOs points to eth1. (I hope this is correct). I have two problems: 1) When I connected to the AP from the XOs, no IP address is generated. 2) We are going to use a hosted XS off site, and I want to configure my box (which is the dhcp server) to point to it. Thanks for the help. Gerald ___ Server-devel mailing list Server-devel@lists.laptop.org http://lists.laptop.org/listinfo/server-devel
Re: [Server-devel] Troubles running F9 mock chroot under F11
2009/9/17 Jerry Vonau jvo...@shaw.ca:
Well not quite the way I would of done it, I was hoping someone else
would test my idea, but glad you got what you needed done. Mind sharing
the yumdownloader routine? I might try to use something like that to
populate my updates repo, then remove any duplicates in the rpms what
would take extra space on the iso.
During the build script:
cat EOF yum.conf
[main]
reposdir=$(pwd)/yumrepos
cachedir=/var/cache/yum-xs
EOF
pushd ${isocopy}/Packages
yumdownloader -c ../../yum.conf --resolve mysql-server mysql php-mysql
expect ImageMagick graphviz php-tidy java-1.6.0-openjdk
java-1.6.0-openjdk-devel
popd
yumrepos/ then contains 2 repo files, one for F9 and one for F9 updates.
The only slightly awkward thing is that the added packages have to be
listed twice, once above and once in the kickstart file. Of course, we
could automate the construction of the ks file but that's adding more
complexity than I'd like to leave behind in Nepal.
Daniel
___
Server-devel mailing list
Server-devel@lists.laptop.org
http://lists.laptop.org/listinfo/server-devel
