| *** khildin has quit IRC | 00:57 | |
| *** menesis has quit IRC | 01:20 | |
| *** th1a has quit IRC | 02:04 | |
| *** yvl has joined #schooltool | 09:02 | |
| *** yvl has quit IRC | 11:55 | |
| *** yvl has joined #schooltool | 11:56 | |
| *** menesis has joined #schooltool | 12:52 | |
| *** khildin has joined #schooltool | 13:11 | |
| *** menesis has quit IRC | 15:40 | |
| *** khildin has quit IRC | 16:57 | |
| *** khildin has joined #schooltool | 17:10 | |
| *** menesis has joined #schooltool | 17:43 | |
| *** replaceafill has joined #schooltool | 18:22 | |
| *** khildin has quit IRC | 20:47 | |
| *** khildin has joined #schooltool | 21:02 | |
| *** th1a has joined #schooltool | 21:22 | |
| *** menesis has quit IRC | 21:23 | |
| *** menesis has joined #schooltool | 22:47 | |
| th1a | hi replaceafill, menesis. | 22:59 |
|---|---|---|
| replaceafill | hi th1a | 23:00 |
| replaceafill | could you give me 5-10 mins | 23:00 |
| replaceafill | i want to show you something | 23:00 |
| th1a | Sure. | 23:00 |
| replaceafill | th1a, http://dev.schooltool.org/primary | 23:02 |
| replaceafill | http://dev.schooltool.org/secondary | 23:03 |
| replaceafill | running zc.zrs | 23:03 |
| replaceafill | any change to primary is shown in secondary | 23:03 |
| replaceafill | after primary goes down, you can change secondary's config to be primary | 23:04 |
| *** khildin has quit IRC | 23:04 | |
| th1a | OK... | 23:08 |
| th1a | So, just to be clear... | 23:08 |
| th1a | We have two ZEO servers? | 23:08 |
| replaceafill | well, yes | 23:08 |
| th1a | I'm not 100% sure I have the basic moving parts down correctly. | 23:08 |
| th1a | So walk me through it. | 23:08 |
| replaceafill | primary, server and client of its own | 23:08 |
| replaceafill | secondary, server (not being used) and client of primary | 23:09 |
| replaceafill | zeo speaking | 23:09 |
| th1a | What's the zeo client? SchoolTool? | 23:09 |
| replaceafill | yes | 23:09 |
| replaceafill | the difference with the default config of SchoolTool is that all of this happens through TCP/IP | 23:10 |
| replaceafill | not regular sockets | 23:10 |
| th1a | So all of SchoolTool's interaction with ZEO changes a bit? | 23:10 |
| replaceafill | the zeoclient config, yes | 23:10 |
| replaceafill | currently | 23:10 |
| replaceafill | all ST instances | 23:10 |
| replaceafill | are zeo servers | 23:11 |
| replaceafill | and clients of their own | 23:11 |
| replaceafill | through sockes | 23:11 |
| replaceafill | sockets* | 23:11 |
| th1a | We'll need a simple diagram of this. | 23:11 |
| replaceafill | :) | 23:11 |
| replaceafill | ok | 23:11 |
| replaceafill | if we try to get this done | 23:11 |
| th1a | OK, let's say "primary" here is the school site. | 23:11 |
| replaceafill | we'll need a way to convert the secondary into primary | 23:12 |
| replaceafill | yes | 23:12 |
| th1a | Then we'd also configure it so that the secondary ZEO instance is running at the central office. | 23:12 |
| th1a | Can that instance be read locally (at the central office)? | 23:13 |
| replaceafill | that's the trick we'll need | 23:13 |
| replaceafill | central office starts as secondary, right? | 23:14 |
| th1a | That's how I've been looking at it. | 23:14 |
| replaceafill | once it's been "sync'ed" | 23:14 |
| replaceafill | you do the "config trick" | 23:14 |
| replaceafill | it becomes a "regular" instance | 23:14 |
| replaceafill | it's own zeo server/client | 23:14 |
| replaceafill | and it can be accessed locally in the central office | 23:15 |
| replaceafill | i'm not sure if this is the best solution | 23:15 |
| replaceafill | but at least, it's *A* solution :) | 23:15 |
| th1a | Well... | 23:15 |
| replaceafill | make some change to primary | 23:15 |
| replaceafill | and i'll take it down | 23:15 |
| replaceafill | so you'll see what happens in secondary | 23:16 |
| replaceafill | it explicitly tells you about a disconnect error | 23:16 |
| th1a | " ZRS also improves scalability since secondary servers provide additional read-only ZEO connections while maintaining their replication functions. ZEO client storages can thus be configured to connect to any available secondary, or load-balance among the available secondary ZRS servers." | 23:17 |
| replaceafill | i thought about setting multiple zeos in the central office instance | 23:18 |
| th1a | So we probably couldn't just hook up regular SchoolTools to these while they were active secondaries, but we could theoretically have some kind of reporting system that queried them in some way. | 23:18 |
| replaceafill | but i didn't try it | 23:18 |
| th1a | We don't have to get too deep into it. | 23:19 |
| replaceafill | ok | 23:19 |
| th1a | Essentially, I think it sounds really good to bring up as an option that is relatively "for free." | 23:19 |
| replaceafill | :) | 23:19 |
| replaceafill | agree | 23:19 |
| replaceafill | thanks ZC :D | 23:20 |
| th1a | Especially since people seem draw to having ALL THE DATA at the central office. | 23:20 |
| th1a | Wait until I tell David Ally. | 23:20 |
| replaceafill | :D | 23:20 |
| replaceafill | well, the good part imho is that we don't need separate reports | 23:20 |
| replaceafill | for central office and schools | 23:21 |
| th1a | But one approach literally might be just writing Python scripts that import SchoolTool objects and extract data for reports. | 23:21 |
| th1a | Oh... | 23:21 |
| th1a | Yes, to a certain extent. | 23:21 |
| replaceafill | we can just write the report | 23:21 |
| th1a | I mean, there would be some different reports. | 23:21 |
| replaceafill | or we could go with python scripts :P | 23:22 |
| th1a | Well... | 23:22 |
| th1a | There are a lot of possibilities. | 23:22 |
| th1a | Depending on what they ACTUALLY NEED. | 23:22 |
| replaceafill | if they want some multi-school reports then python scripts should be the way to go | 23:23 |
| th1a | I assume if you'd just connect the regular SchoolTool interface to a read only ZEO, you'd get fouled up pretty wuick. | 23:23 |
| replaceafill | i'd like to experiment with that | 23:23 |
| replaceafill | having several primaries | 23:24 |
| replaceafill | (with their own secondary serving zeos) | 23:24 |
| replaceafill | and try to get data from the secondaries | 23:24 |
| replaceafill | list of people or something | 23:24 |
| replaceafill | something simple | 23:24 |
| th1a | Yeah, that would be worth knowing if it works. | 23:24 |
| replaceafill | +1 | 23:24 |
| th1a | Because I'll probably end up proposing it as a possible solution. | 23:25 |
| replaceafill | my only concern is | 23:25 |
| replaceafill | about tcp/ip access | 23:25 |
| replaceafill | the central office should have access to the local site | 23:25 |
| replaceafill | i see it more as a pull action | 23:26 |
| replaceafill | than a push from the school | 23:26 |
| th1a | It isn't entirely practical. | 23:26 |
| replaceafill | imagine blocked ports and stuff | 23:26 |
| th1a | What is it by default? | 23:26 |
| replaceafill | uh? | 23:26 |
| th1a | Pull or push? | 23:27 |
| replaceafill | not sure, but i imagine pull from the zeo client | 23:28 |
| th1a | The whole idea is a little impractical, but they might like it. | 23:28 |
| replaceafill | i'd like to know what zookeeper is :) | 23:29 |
| replaceafill | it's in the zrs component.xml | 23:29 |
| replaceafill | and it's briefly mentioned in the docs | 23:29 |
| th1a | This is sort of the fallback if their current data warehouse only understands undocumented blobs from Capita/MIS. | 23:29 |
| replaceafill | so, should i try to make the multi-clients report? | 23:30 |
| th1a | Yeah. It is worth a few hours. | 23:30 |
| replaceafill | kk | 23:31 |
| th1a | https://pypi.python.org/pypi/zc.zk/0.5.2 | 23:31 |
| th1a | It is actually an Apache thing. | 23:32 |
| replaceafill | yeah | 23:32 |
| replaceafill | i found their website | 23:32 |
| th1a | I couldn't imagine that you hadn't. ;-) | 23:32 |
| replaceafill | :D | 23:32 |
| th1a | OK. | 23:32 |
| th1a | menesis, ayt? | 23:33 |
| replaceafill | i won't go that deep though :P | 23:33 |
| th1a | So I'm going to start on the RFQ in earnest tomorrow. | 23:33 |
| th1a | Did you look at the student finance piece? | 23:33 |
| th1a | I mean work on it over the weekend? | 23:34 |
| replaceafill | well, kind of | 23:34 |
| replaceafill | didn't get too far | 23:34 |
| replaceafill | interfaces and stuff | 23:34 |
| replaceafill | nothing working | 23:34 |
| replaceafill | but their email helped me | 23:35 |
| th1a | Sure. | 23:35 |
| replaceafill | doesn't look that scary | 23:35 |
| th1a | No. | 23:35 |
| th1a | We should try to get back to release cleanup, though. | 23:36 |
| replaceafill | ok | 23:36 |
| th1a | We have two weeks, and then we'll move on. | 23:36 |
| replaceafill | two weeks of bug fixing? | 23:37 |
| th1a | Yeah. | 23:37 |
| replaceafill | ah ok | 23:37 |
| th1a | Start on the ARK stories after Easter. | 23:37 |
| replaceafill | one thing i did during the weekend is that i went back to the past | 23:37 |
| replaceafill | using bzr | 23:37 |
| replaceafill | while i was working on the sakeji report | 23:38 |
| replaceafill | i started noticing some gradebook methods requiring things the gradebook already knows | 23:38 |
| replaceafill | like worksheets | 23:38 |
| replaceafill | and i started wtf?!? | 23:38 |
| replaceafill | so i started replaying the gradebook bzr history :D | 23:38 |
| replaceafill | and i realized when some changes were done | 23:39 |
| replaceafill | and at some extent, why | 23:39 |
| replaceafill | oh | 23:39 |
| th1a | Ah. | 23:39 |
| replaceafill | and i tried fixing the selenium/chrome issue | 23:39 |
| replaceafill | but i failed :( | 23:40 |
| th1a | So are those things that can be unwound or is it more of a "oh, that's the reason" situation. | 23:40 |
| replaceafill | well, we just need to update the code | 23:40 |
| replaceafill | the code started working in some way | 23:40 |
| replaceafill | was changed in some parts | 23:40 |
| replaceafill | but not all | 23:40 |
| replaceafill | and when you see those unchanged parts | 23:41 |
| replaceafill | you get confused | 23:41 |
| replaceafill | but it's not a big deal | 23:41 |
| replaceafill | we can move on | 23:41 |
| replaceafill | the selenium thing is more worth to fix | 23:41 |
| replaceafill | but only affects downloads at this point | 23:42 |
| th1a | OK. Well, it seems like a lot of the small fixes/mini-features are in the gradebook. | 23:42 |
| replaceafill | yes | 23:42 |
| th1a | But generally, the bugs are identified and prioritized, so just go ahead. | 23:43 |
| replaceafill | kk | 23:43 |
| replaceafill | i have some fixes that i need to commit | 23:43 |
| replaceafill | i'll update the bugs accordingly | 23:43 |
| replaceafill | i'll keep working on zrs just for today | 23:43 |
| th1a | Oh, I guess we want to do bugs probably until Palm Sunday, not Easter. | 23:43 |
| th1a | That's two weeks of work. | 23:44 |
| replaceafill | if no external request comes ;) we should do a lot of fixing | 23:44 |
| th1a | That gives us two weeks head start in theory, but probably not at all. | 23:44 |
| th1a | Probably a little. | 23:44 |
| replaceafill | did you see jelkner's email? | 23:45 |
| th1a | Oh, shit. Thanks for reminding me. | 23:46 |
| replaceafill | :D | 23:46 |
| th1a | He sent that friday night or something. | 23:46 |
| th1a | OK, replying now. | 23:47 |
| replaceafill | oh now that you mention hours! | 23:47 |
| replaceafill | have you changed your time yet? | 23:47 |
| replaceafill | +1 hour? | 23:47 |
| th1a | If menesis isn't here we're pretty much done. | 23:47 |
| th1a | Yeah, we did. | 23:47 |
| th1a | Can we shift an hour earlier? | 23:47 |
| replaceafill | so, we're meeting 9 pm your time now? | 23:47 |
| replaceafill | ah ok | 23:47 |
| replaceafill | sure | 23:47 |
| replaceafill | 8 pm your time, correct? | 23:47 |
| replaceafill | that'll be 1 pm for me | 23:47 |
| th1a | OK, yeah. | 23:48 |
| th1a | That's right. | 23:48 |
| replaceafill | ok, thanks | 23:48 |
| replaceafill | that's it from me th1a | 23:48 |
| th1a | OK, thanks replaceafill. I'll keep you posted on the bid. I guess we have two weeks for that as well. | 23:49 |
| * th1a drops the bag of gravel. | 23:49 | |
Generated by irclog2html.py 2.15.1 by Marius Gedminas - find it at mg.pov.lt!