« Previous day | Index | Today | Next day » Channels | Search | Join
All times shown according to UTC.
| Time | Nick | Message |
|---|---|---|
| 00:28 | fm <fm!~Paraguay mail.paraguayeduca.org> has joined #olpc-paraguay |
|
| 03:26 | tch has left #olpc-paraguay | |
| 05:02 | walterbender <walterbender!~chatzilla201.192.97.46> has joined #olpc-paraguay |
|
| 05:20 | walterbender has quit IRC | |
| 10:30 | Cynthiar <Cynthiar!~Paraguaymail.paraguayeduca.org> has joined #olpc-paraguay |
|
| 10:32 | Cynthiar | Buen Día Jorge está por ahi? |
| 10:35 | Cynthiar has quit IRC | |
| 11:19 | javix <javix!~javix200.61.226.194> has joined #olpc-paraguay |
|
| 12:07 | car|0s <car|0s!~carlosmail.paraguayeduca.org> has joined #olpc-paraguay |
|
| 12:29 | fm <fm!~Paraguaymail.paraguayeduca.org> has joined #olpc-paraguay |
|
| 12:32 | jasg <jasg!~jorgemail.paraguayeduca.org> has joined #olpc-paraguay |
|
| 12:40 | walterbender <walterbender!~chatzilla201.192.97.46> has joined #olpc-paraguay |
|
| 12:52 | tch <tch!~tch186.16.74.160> has joined #olpc-paraguay |
|
| 12:57 | rgs__ | car|0s: Oct 6 08:45:41 schoolserver puppetd[3777]: Finished catalog run in 226.94 secon |
| 12:58 | car|0s: puppet corrio en la 278 esta matina | |
| 12:58 | car|0s: ahora veo pq no copio los leases | |
| 12:58 | car|0s | rgs__, si, pero no recibio los leases |
| 12:58 | rgs__, y de acuerdo a mothership, estos fueron enviados | |
| 12:58 | rgs__ | car|0s: la pregunta es si los leases se generaron en mothership entonces |
| 12:58 | car|0s | rgs__, si, se generaron |
| 13:00 | Oct 6 08:37:05 schoolserver puppetd[2856]: XMLRPC returned wrong size. Retrying. | |
| 13:00 | rgs__, ^^ | |
| 13:00 | rgs__ | car|0s: q interesante |
| 13:00 | car|0s | rgs__, el unico error q encontre |
| 13:00 | rgs__ | car|0s: mirando |
| 13:02 | car|0s: ese error viste en /var/log/messages de la 278? | |
| 13:02 | car|0s | rgs__, si |
| 13:02 | Oct 6 08:30:56 schoolserver puppetd[1802]: (//Node[base]/leases/File[/var/lib/xsleases/leases.dat]) Filebucketed to with sum ac18d1f9a6fbffe9a424200de9510258 | |
| 13:02 | Oct 6 08:30:56 schoolserver puppetd[1802]: (//Node[base]/leases/File[/var/lib/xsleases/leases.dat]/source) replacing from source puppet:///leases/schoolserver.escuela278.caacupe.paraguayeduca.org with contents {md5}b03842920a08b63f8b2419bed1fbd8b6 | |
| 13:08 | rgs__, la 278 recibio el leases.dat nuevo | |
| 13:09 | rgs__ | car|0s: pero? |
| 13:09 | car|0s | rgs__, encontre en /var/lib/xsleases/leases.dat |
| 13:09 | rgs__ | car|0s: no proceso eso? |
| 13:09 | car|0s | rgs__, y.. eso estoy mirando.. |
| 13:09 | rgs__ | car|0s: dp corre un script que parte ese .dat en multiples leases files |
| 13:09 | car|0s | puppet-xs-lease-import |
| 13:09 | rgs__ | car|0s: veo que hay leases q se crearon hoy |
| 13:09 | car|0s | q a su vez llama a otro script |
| 13:09 | rgs__ | car|0s: espera |
| 13:10 | car|0s: yo entre a /library/xs-activation-server | |
| 13:10 | car|0s: y corri find -mtime -1 para ver q archivos se crearon en las ultimas 24hrs | |
| 13:10 | y me salieron una partida | |
| 13:10 | osea que finalmente corrio parece | |
| 13:11 | car|0s | rgs__, asi parece.. no veo mas el millon de request de leases de las xo en messages |
| 13:11 | rgs__, pero yo corri puppet-xs-lease-import | |
| 13:12 | rgs__ | car|0s: hmmm... hace cuanto corriste? |
| 13:12 | car|0s | rgs__, ni 5 minutos |
| 13:12 | rgs__ | car|0s: una pregunta , q hora es en Asu? |
| 13:12 | car|0s: las 10am vd? | |
| 13:13 | car|0s | rgs__, asi es |
| 13:13 | rgs__ | car|0s: pq el server dice 9:12 |
| 13:13 | car|0s: comentario al margen nomas... no funciona el horario de verano | |
| 13:13 | car|0s | rgs__, pq hubo cambio de hora... |
| 13:13 | rgs__ | car|0s: pero ya sabismo de este problema, el tema es que son un fedora del anho del pedo con timezones viejos |
| 13:13 | car|0s | rgs__, sip.. encima el nuestro tan particular |
| 13:13 | rgs__ | car|0s: volviendo a lo de los leases.. porque sera q puppet no corre el script? |
| 13:14 | car|0s | rgs__, voy a mirar la 254.. me reportaron el mismo problema |
| 13:15 | rgs__, schoolserver.escuela40.caacupe.paraguayeduca.org | |
| 13:15 | rgs__, 254 offline | |
| 13:16 | walterbender has quit IRC | |
| 13:16 | rralcala <rralcala!~rralcalamail.paraguayeduca.org> has joined #olpc-paraguay |
|
| 13:18 | rralcala has quit IRC | |
| 13:19 | rralcala <rralcala!~rralcalamail.paraguayeduca.org> has joined #olpc-paraguay |
|
| 13:20 | rgs__ | rralcala: el Alki |
| 13:20 | car|0s: con respecto a Puppet | |
| 13:20 | car|0s: 1) Puppetmaster deberia correr sobre apache/passenger | |
| 13:20 | bernie: ^ | |
| 13:20 | car|0s: 2) un update de los puppet clients seria interesante | |
| 13:21 | tch | rralcala: que tal caacupe ayer? |
| 13:21 | rralcala | tch: tranquilo |
| 13:32 | rgs__ | bernie: around? |
| 13:32 | car|0s | Oct 6 09:31:37 schoolserver smartd[3618]: Device: /dev/sda, 291 Currently unreadable (pending) sectors |
| 13:32 | Oct 6 09:31:37 schoolserver smartd[3618]: Device: /dev/sda, 291 Offline uncorrectable sectors | |
| 13:32 | rgs__ | car|0s: ping |
| 13:32 | car|0s | la gran.. otro disco duro q va morir |
| 13:33 | rgs__ | car|0s: ouch |
| 13:33 | car|0s: vos podes commitear cambios al site.pp config vd? | |
| 13:33 | car|0s | rgs__, eso era para el puppet, no? |
| 13:33 | rgs__ | car|0s: usando nuestro repo interno de svn para cosas privadas? |
| 13:33 | car|0s | rgs__, 1 sola vez hice contigo |
| 13:34 | rgs__ | car|0s: pero sos muy inteligente, una sola vez nomas hace falta mostrarte las cosas :) |
| 13:34 | car|0s: tenes a mano el directorio puppetconfigs? | |
| 13:34 | car|0s: tenemos q agregar el pub key de rralcala | |
| 13:34 | car|0s: pa que pueda loggearse a los schoolservers | |
| 13:35 | car|0s | rgs__, la verdad q no recuerdo.. tengo q revisar mis notas |
| 13:35 | rgs__ | car|0s: ?? |
| 13:35 | car|0s: no te preocupes, yo te guio | |
| 13:35 | car|0s: encontra nomas la carpeta puppetconfigs | |
| 13:35 | car|0s: q es un checkout del repo | |
| 13:36 | car|0s: y ahi adentro hay un archivo site.pp | |
| 13:36 | car|0s | Oct 6 09:31:54 schoolserver puppetd[10062]: Could not call puppetmaster.getconfig: #<#<Class:0xb7a0717c>: execution expired> |
| 13:36 | Oct 6 09:31:54 schoolserver puppetd[10062]: Could not retrieve catalog: execution expired | |
| 13:36 | Oct 6 09:31:57 schoolserver puppetd[10062]: Starting catalog run | |
| 13:36 | rgs__ | car|0s: encontraste? |
| 13:37 | car|0s | rgs__, si, estoy ahi |
| 13:38 | rgs__ | car|0s: svn update |
| 13:38 | car|0s: pa asegurarnos que tenes lo mas fresco | |
| 13:38 | y luego necesito que edites site.pp | |
| 13:38 | bernie | rgs__: pong |
| 13:38 | car|0s | U site.pp |
| 13:38 | Actualizado a la revisión 150. | |
| 13:38 | bernie | rralcala: one of the first tasks to do on the pyedu infrastructure is probably to update all machines to a recent version of fedora |
| 13:39 | rgs__ | bernie: NO |
| 13:39 | bernie | rralcala: I did a few of them several months ago |
| 13:39 | rgs__: because it would cause breakage? | |
| 13:39 | rgs__ | bernie: lets focus on giving him the tools to solve the actual real problems he has now, other wise we'll overwhelm him |
| 13:39 | bernie: because I'd leave that for week 2 | |
| 13:39 | car|0s | rgs__, bernie paofila for sure will break |
| 13:40 | rgs__ | car|0s: podrias fijarte en el site.pp como estan declarados los usuarios |
| 13:40 | car|0s: y agregar uno para rralcala con su pub key | |
| 13:40 | bernie: car|0s will add rralcala to site.pp so he is user along with his pub key is created in each school server | |
| 13:40 | car|0s | rralcala, me podes pegar tu llave? |
| 13:41 | bernie | car|0s: sure, many things will break, that's to be expected. Though it's probably better to do it until rgs and I are around to help fix the inevitable fallout. |
| 13:42 | rgs__: yes, seems reasonable | |
| 13:42 | rgs__ | bernie: universe might freeze first |
| 13:42 | rralcala | ssh-rsa AAAAB3NzaC1yc2EAAAABIwAAAQEA0NVCgHq3nI2rXb06qiq/xKKdK86tHqmxTqwXTzmtGsQjfNdGsiFs3/kaWT2lkn4Fc1kKf7eVBTbf4Myn5zKMR8P5+j/GBWLd+Ze1p6hOrfYS+9+yzp4i91uOBUkdTgUPvp2HGkOGT9BDa8SYlK9CjlSsaCGcYTcMFR4eXcpBO52vxf5lfnDWEB/8+0sfMkDXwMi9JAIcTxkOC7bLZQp7pO+cTDivsoTZ/P5EFbnaSCktsvKP7vAyNxUVFQFlsaHSepars6VdgczHKGHqMxwzzo23XJ5EhLKwuj//yWEDMv0s6ntOcIrI1rj94E8Gjn3IUyVuuDVa4D2jP575T7TUGQ== |
| 13:43 | rgs__ | car|0s: cuando hayas editado site.pp, pega en pastebin.org el output de svn diff si queres q mire |
| 13:43 | car|0s | rgs__, ok |
| 13:48 | morganya_afk <morganya_afk!~chatzillamail.paraguayeduca.org> has joined #olpc-paraguay |
|
| 13:48 | morganya_afk is now known as morganya | |
| 13:49 | rgs__ | car|0s: ya? |
| 13:49 | fm has quit IRC | |
| 13:51 | car|0s | rgs__, si, ya esta |
| 13:51 | rgs__, el pastebin lo q tarda.. nuestra super conexion | |
| 13:52 | rgs__, http://pastebin.org/59353 | |
| 13:52 | rgs__ | car|0s: un sec |
| 13:58 | car|0s: https://wiki-interno.paraguaye[…]CaacupeMothership | |
| 14:00 | car|0s: no vayas a borrar el usuario de dsd | |
| 14:00 | car|0s: agrega lineas nuevas para el de rralcala | |
| 14:01 | car|0s | rgs__, ok. no tengo q modificar sudoers tb? |
| 14:01 | rgs__ | car|0s: huh? |
| 14:07 | car|0s | rgs__, http://pastebin.org/59548 |
| 14:08 | rgs__ | car|0s: tambien falta la seccion donde se crea el usuario rralcala |
| 14:10 | car|0s | rgs__, http://pastebin.org/59575 |
| 14:11 | rgs__ | car|0s: me gusta , valida con rralcala para ver si a el le gusta tb |
| 14:11 | car|0s | rralcala, http://pastebin.org/59575 |
| 14:12 | rgs__, dp svn commit y listo? | |
| 14:15 | walterbender <walterbender!~chatzilla201.192.97.46> has joined #olpc-paraguay |
|
| 14:19 | morganya is now known as morganya_afk | |
| 14:20 | rgs__ | car|0s: sip |
| 14:20 | car|0s: rralcala se encarga del resto | |
| 14:20 | rralcala: ^ | |
| 14:25 | car|0s: hay que restart-ear los puppets | |
| 14:25 | pa que se cree el usuario de rralcala en cada server | |
| 14:26 | car|0s | rgs__, rralcala algun servidor en particular o quieren todos? |
| 14:26 | rgs__, rralcala eventualemente puppet va correr en todos (cronjob) | |
| 14:27 | rgs__ | car|0s: metele en todos - asi ganamos tiempo |
| 14:28 | walterbender has quit IRC | |
| 14:29 | morganya_afk is now known as morganya | |
| 14:43 | morganya | rgs__: can the uid in a Journal record be decoded into a name? It looks like a hash ... |
| 14:45 | rgs__: actually, it looks like the record is saved by the machine's serial number ... that would do too. | |
| 14:46 | rgs__ | morganya: what do you want to do with the uid |
| 14:47 | morganya: i think its more of an ID (actually uid = unique identifier) | |
| 14:47 | morganya: which is probably used to distinguish each entry in a unambigious way | |
| 14:47 | morganya | rgs__: I was curious about decoding to gender and looking for differences in use by gender (I have a script that categorizes names) |
| 14:48 | rgs__ | morganya: so you want something like getName(UID) ? |
| 14:48 | morganya | rgs__: ah, then the serial number would be best. I have the table of those from inventario |
| 14:48 | rgs__ | morganya: yup |
| 14:48 | morganya | getName(serial) would be fine too |
| 14:48 | rgs__ | morganya: not sure the name configured for the XO or the SN can be deduced from the UID or any other key in the meta data |
| 14:48 | morganya | as long as serial is also scraped along with the metadata |
| 14:48 | rgs__ | morganya: BUT |
| 14:49 | morganya | rgs__: (serial is the directory name for the backup/metadata?) |
| 14:49 | rgs__ | morganya: we know the SN because the backup is associated to a SN |
| 14:49 | morganya | yeah |
| 14:49 | rgs__ | morganya: yes |
| 14:49 | morganya | perfect :) |
| 14:49 | rgs__ | morganya: so what we could do is |
| 14:49 | morganya | rgs__: we can also determine whether they're in school in the morning or afternoon from that |
| 14:50 | rgs__ | morganya: mix your code (that looks up gender based on SN) with the rest of the analysis code |
| 14:50 | morganya: once you have the SN you can do many fancy things :) | |
| 14:50 | morganya | rgs__: and differentiate between activities used in class vs. activities at home |
| 14:50 | rgs__: yeah :) | |
| 14:50 | rgs__ | morganya: niiiece |
| 14:50 | morganya: would you be so kind to dump these ideas into the Wiki article we starte? | |
| 14:50 | morganya | sure |
| 14:51 | rgs__ | morganya: having this publicly available will lower the barriers so can convert them into code |
| 14:51 | morganya: dont forget to register/login before editing the wiki so we know who edited what | |
| 14:51 | morganya | sounds good |
| 14:51 | [scs] | rralcala: http://www.llibrewebsantillana.cat/ |
| 14:55 | rgs__ | morganya: (and then reply to my 1st email with a heads up - so that everyone can know we are making some progress) |
| 14:56 | morganya | rgs__: sure :) |
| 15:02 | rgs__: we don't have duration, do we? The amount of time kids spent on the activity? Or is that the timestamp? | |
| 15:04 | rgs__ | morganya: we have mtime and timestamp |
| 15:05 | morganya: if want is the start and the other when the last edit occured - that might be useful | |
| 15:05 | morganya: the interesting thing of this whole discussion is that stuff like this (the need to know for how a long an activity was used - or at least open) can then (more or less easily) be implement | |
| 15:06 | ed | |
| 15:06 | morganya | rgs__: yeah |
| 15:06 | rgs__ | morganya: knowing how long an activity was used is veeery important. And we could even think of crazy heuristics to figure out the intensity of the use |
| 15:07 | morganya | rgs__: we don't have the number of times an activity instance has been opened, either ... that would be another that sounds like it should be easy to add and would be interesting. |
| 15:07 | rgs__: so we could add the total amount of time an activity was used and the number of separate occasions it was opened (so was this work done all at once or spread out) | |
| 15:08 | rgs__: hmm ... for intensity, you could record the number of clicks or mouse movements, but that starts to get kind of creepy. I want to respect the kids' privacy, too ... | |
| 15:08 | rgs__: ... even though the social scientist in me wants as much data as possible for analysis. ;) | |
| 15:10 | rgs__ | morganya: yes - I can see the title of Malcom Gladwell's next book |
| 15:11 | morganya: re number of times an activity was opened, for each time an activity is open you get a Journal entry so... | |
| 15:11 | morganya | rgs__: *laugh* I just have to learn to *write* like him first ... then we'll be set ;) |
| 15:11 | rgs__: ah, I see -- they already generate separate entries | |
| 15:12 | rgs__ | morganya: activity opened --> entry created |
| 15:12 | morganya | rgs__: so we could know from the title that it's the same instance of an activity opened again, unless they change it |
| 15:12 | rgs__ | (--> == implies) |
| 15:12 | which begs the questin.. does it make sense to save an empty entry? | |
| 15:13 | I am sure it has been discussed ad nauseam - bernie or walter should be able to elaborate more on the rationale of creating empty entries | |
| 15:13 | morganya | how would the entry be empty? If something is opened and then closed again right away, for example? |
| 15:13 | rgs__ | morganya: you can know from the activity_id I think |
| 15:14 | morganya | rgs__: ah, that persists ... perfect |
| 15:14 | rgs__ | morganya: you can always now the correspondance between an entry and an activity irregardless of if a title was set or not |
| 15:15 | morganya | rgs__: it isn't the case that all instances of Navegar, for example, have the same activity ID, even if they are new sessions of the browser (rather than starting from previous work)? |
| 15:15 | rgs__ | [scs]: ping |
| 15:15 | morganya | rgs__: for example, if I open two instances of Navegar at the same time, will they have unique activity_ids? |
| 15:16 | rgs__ | morganya: no idea about that - I'd have to check |
| 15:27 | rralcala has quit IRC | |
| 15:29 | morganya_ <morganya_!~chatzillamail.paraguayeduca.org> has joined #olpc-paraguay |
|
| 15:30 | morganya has quit IRC | |
| 15:30 | morganya_ is now known as morganya | |
| 15:32 | rgs__ | morganya: around_ |
| 15:32 | ? | |
| 15:32 | morganya: a couple of notes of your latest additions | |
| 15:33 | morganya | rgs__: my description of the extra metadata is a bit vague right now, sorry :) |
| 15:33 | rgs__: sure | |
| 15:33 | rgs__: I can add more detail to that one section later, though | |
| 15:33 | rgs__ | do you have a table that classifies games (i.e.: educational vs non-educational) ? We'd probably have to build such table |
| 15:34 | the question about how activities spread is an interesting one - I wonder how we could tackle it | |
| 15:34 | morganya | rgs__: yes, we'd have to build it. I have a decent sense of what games are *considered* educational vs. non-educational by teachers and parents. :) |
| 15:34 | rgs__ | morganya: that is a great start! :-) |
| 15:35 | morganya: so we have 2 milestones ahead - that I can think of - right now: | |
| 15:36 | morganya | rgs__: I was thinking that we could look for the first instance of activity use for each person (disregarding the most common activities) to determine when an activity was spread to that person ... but how to establish connections would be more difficult. There could be some weighted function based on school/grade/class ... |
| 15:36 | rgs__ | 1) figuring out how to measure the things that right now can't be measured (i.e.: usage time of an activity) |
| 15:36 | (this requires design, code, testing and installation - very time consuming) | |
| 15:37 | 2) crunch the journals into a DB or something from which we can run all the questions | |
| 15:37 | morganya | rgs__: this would be modifying the Journal metadata that is saved? |
| 15:37 | rgs__: they're just dumped into a text file now, right? | |
| 15:37 | rgs__ | morganya: what time frame are we looking into? Bear in mind that many things will probably take a lot of time. You are opening a new path for future investigations though! |
| 15:37 | morganya: yes and yes | |
| 15:38 | morganya: I suspect the changes we are after are non invasive | |
| 15:38 | morganya: none the lest you'd have to patch 10K Sugar installations | |
| 15:38 | morganya: not impossible but not trivial either | |
| 15:38 | morganya: plus i think bernie has turned off the OS update mechanism | |
| 15:38 | morganya: which would have been an interesting vector to spread out the changes | |
| 15:39 | morganya | rgs__: hmm ... well, I feel like this could be really interesting, so I'm willing to put time toward it ... my Uruguay time keeps shrinking more and more, though. :) My other main task is to write and test an exam for third and sixth grades. I have a draft already, but it needs translation and validation ... |
| 15:39 | rgs__ | bernie: olpc-update disabled on the 0.84 builds used in Caacupe? Have about Dextrose? |
| 15:39 | tch: ^ | |
| 15:39 | car|0s: ^ | |
| 15:39 | morganya | rgs__: bbl |
| 15:39 | tch | rgs_: hasta donde se, no tenemos corriendo eso, y si tenemos jamas se probo |
| 15:40 | rgs__ | morganya: we should have 2 parts: 1) do whatever we can with the available data now and 2) plan for the changes needed to get a richer data set |
| 15:40 | tch: osea que si queres sacar un update para el datastore seria: | |
| 15:40 | 1) re-flash (tambien necesitas un nuevo build) | |
| 15:40 | 2) rpm -Uvh a mano 10K veces? | |
| 15:41 | tch | basicamente, no llegamos a cubrir la parte de updates para dx1 |
| 15:43 | rgs__ | tch: no entiendo eso? |
| 15:43 | tch: hubo una discussion de como se harian los updates automaticos encima de dx1 pero no se concluyo nada? | |
| 15:43 | tch: tipo olpc-update vs yum upgrade etc? | |
| 15:44 | tch | si, yo me tire por lo de yum pero no se llego a implementar |
| 15:44 | mucho overhead | |
| 15:44 | rgs__ | tch: pero hacia falta un breakthrough para meterle pure yum upgrades? |
| 15:44 | tch | en general, se tuvo que en invertir en otras cuestiones |
| 15:44 | rgs__ | tch: entiendo |
| 15:44 | morganya is now known as morganya_afk | |
| 15:45 | rgs__ | tch: bottom line: no hay un network based (transparent) update scheme hoy en dia? |
| 15:45 | tch | nope, todavia no llego eso a la realidad, |
| 15:46 | ni siquiera probamos el olpc-update | |
| 15:47 | rgs__ | tch: ok |
| 15:48 | tch | ademas, dextrose se esta usando en 60 maquinas nomas hoy en dia, |
| 15:48 | hasta que se le de el "go" al resto de DO | |
| 15:53 | rgs__ | morganya_afk: so, there is probably a big Software update (migrating from F11/Sugar-0.84 to Dextrose-1) ahead. If we come up with trivial non-intrusive patches to the Journal in order to have it help us answer the questions we are pursuing we might be able to piggy back on those efforts and have every laptop in Caacupe loaded with a version of Sugar that is friendly to number crunching |
| 15:54 | morganya_afk: the bad thing is that this means the data wouldn't be able before April/May '11 | |
| 16:01 | s/able/available/g | |
| 16:05 | jasg has quit IRC | |
| 16:59 | bernie | rgs__: olpc-update is enabled (but untested) in dextrose-1 and dextrose-2 |
| 16:59 | rgs__: not in dextrose-0 (aka f11-xo1-py) | |
| 17:00 | rgs__ | bernie: ah |
| 17:00 | bernie | rgs__: nobody tested it though. I would recommend against trying to use it for updates. |
| 17:01 | rgs__: the easiest and fastest method remains handing over usb sticks to teachers and children | |
| 17:01 | rgs__: in uruguay they told me they just publish the image on their web site | |
| 17:01 | rgs__: then teachers and volunteers do the rest | |
| 17:02 | rgs__ | bernie: nice |
| 17:03 | bernie | rgs__: I asked: "what is some school does not upgrade?" the candid answer: "if they don't upgrade, then they keep running the old version!" |
| 17:04 | rgs__ | bernie: Pacita's universe doesn't seem to follow those rules |
| 17:22 | morganya_afk is now known as morganya | |
| 17:22 | morganya | rgs__: back :) |
| 17:23 | rgs__ | morganya: read the back log , I have to go out in a bit |
| 17:23 | morganya | rgs__: April-May 2011 is better than nothing. :) And we can do what we can with the data now too. |
| 17:23 | rgs__: no prob! sorry I had to run for a while too ... had to go with Carla | |
| 17:27 | rgs__ | morganya: the 2 phase approach sounds nice |
| 17:46 | rralcala <rralcala!~rralcalamail.paraguayeduca.org> has joined #olpc-paraguay |
|
| 17:46 | car|0s | rralcala, http://pastebin.org/62467 |
| 17:47 | jasg <jasg!~jorgemail.paraguayeduca.org> has joined #olpc-paraguay |
|
| 17:52 | car|0s | rralcala, http://pastebin.org/62547 el squid |
| 17:59 | carlos__ <carlos__!~carlos170.51.37.39> has joined #olpc-paraguay |
|
| 18:03 | car|0s has quit IRC | |
| 18:23 | tch | From 0.90 release notes: "Protected activities: The GConf key \"/schemas/desktop/sugar/protected_activities\" has been added. You can set a list of activities that users will not be allowed to erase through the list view." nice nice |
| 18:35 | dfarning_ <dfarning_!~dfarning71.13.116.9> has joined #olpc-paraguay |
|
| 18:53 | carlos__ | tch, q grande letinche. fue aceptado el patch upstream? |
| 18:53 | tch | sehh el mes pasado ya, |
| 18:53 | me da risa nomas porque pense que no iban a publicar jaja | |
| 18:54 | porque viste que es medio controversial | |
| 18:54 | carlos__ | tch, depende el punto de vista.. |
| 19:18 | morganya | rgs__: just sent my code ... forgive the hackiness of it (I'm kind of out of practice ;)) |
| 19:18 | rgs__: I'm also curious about the internet logs, but we can talk about those later ... just want to keep it on the table :) | |
| 19:20 | rgs__ | morganya: we've got scs blessing to access though |
| 19:20 | morganya | rgs__: yep |
| 19:20 | rgs__ | morganya: the technical challenge the remains is go to co-relate those with SNs |
| 19:20 | (again - we need to know stuff like gender, age, moring/afternoon shift) | |
| 19:21 | morganya | rgs__: yes ... I can add the necessary function calls to the classes in the code I sent |
| 19:21 | rgs__: that part will be easy :) | |
| 19:21 | rgs__ | so you can come up with Gladcom-ish facts like "why do kids in the afternoon perfer to access directly Wikipedia instead of using Google as a gateway?" |
| 19:21 | is it because they have lots of meat on lunch? | |
| 19:21 | morganya | haha :) |
| 19:21 | rgs__ | morganya: not so fast! |
| 19:22 | morganya: unless you've figured out how to match Squid logs with SN info | |
| 19:22 | morganya: I know you've figured out getSomeAttribute(SN) | |
| 19:22 | morganya | rgs__: though the way it works now is to keep the whole data structure in memory ... the number of problems or laptops is small enough now. but just a warning ... this probably shouldn't be run in Uruguay (though they probably don't use Inventario anyway) ;) |
| 19:22 | rgs__ | morganya: but in the case of internet logs there is no SN associated - only a transient IP address |
| 19:23 | morganya | ah yes |
| 19:23 | rgs__ | morganya: so, we need a trick to fix that inconvenience |
| 19:23 | morganya: (so hopefully - we can rebuilt the info from what went on in the past months) | |
| 19:23 | morganya | rgs__: I was referring more to the journals ... but yeah, the server logs won't have nearly as much metadata |
| 19:23 | rgs__ | morganya: otherwise we'll have to re-configure stuff and wait a few more months until the new data gets generate :( |
| 19:24 | morganya: ahhh.. you want to come learn about the Internet browsing habits from the Journal entries... | |
| 19:24 | morganya | oh, one possible field for a "Navegar" instance in the future would be the last URL accessed ... or some URL record. Last accessed seems more useful than first accessed. |
| 19:24 | rgs__ | morganya: that might work.. because Browse persists the surfing history |
| 19:25 | morganya: actually , learned this from my time hacking Browse back in July, for each instance of Browse the whole Browsing history is saved | |
| 19:25 | morganya | rgs__: really! |
| 19:25 | rgs__: wow ... | |
| 19:26 | rgs__: if kids clear their journals, do those entries still persist on the school server? or are they erased to match the state of the machine on the next backup? | |
| 19:27 | rgs__: also, what percentage of kids have backups on the school servers? Is it all of them? | |
| 19:27 | rgs__ | re #1 |
| 19:27 | morganya | rgs__: or do they have to do something special to trigger a backup? |
| 19:27 | rgs__: sorry, a flood of questions! :) | |
| 19:28 | rgs__ | there are multiple snapshots -although I still havent figured out correctly what the backup strategy really covers |
| 19:28 | re #2 | |
| 19:28 | after the massive update to F11/0.84 it seems many laptops weren't re-registered against their corresponding school server - hence disabling effectively the backup mech | |
| 19:28 | re #3 | |
| 19:29 | the only pre-requisite to have backups running is being registered against a School Server | |
| 19:29 | morganya | great, thanks :) |
| 19:33 | rgs__: we can also check how the coverage is against the full list of laptops, though we won't know if any particular laptop's logs are incomplete (e.g. is that month gap because the kid was clearing her journal, because the laptop wasn't registering with the server, or because she didn't use her laptop for a month?) | |
| 19:34 | rgs__ | morganya: gotta run (again!). Can we resume this later? |
| 19:34 | morganya | rgs__: no prob! |
| 20:15 | carlos__ has quit IRC | |
| 20:24 | dfarning_ has quit IRC | |
| 20:33 | tch has quit IRC | |
| 20:59 | rralcala has quit IRC | |
| 21:05 | tch <tch!~tch186.16.74.160> has joined #olpc-paraguay |
|
| 21:48 | jasg has quit IRC | |
| 21:52 | morganya has quit IRC | |
| 22:12 | rgs__ | [scs]: ping |
| 22:40 | [scs] | rgs__: pong |
| 22:46 | rgs__ | [scs]: Profe viste el tema del pasaje de Walter con Leti? |
| 23:44 | walterbender <walterbender!~chatzilla201.192.97.46> has joined #olpc-paraguay |
« Previous day | Index | Today | Next day » Channels | Search | Join