*** Joins: giallu (~giallu@fedora/giallu) | 00:32 | |
*** Joins: Paul_46 (~IceChat09@cpc1-enfi15-2-0-cust580.hari.cable.virginmedia.com) | 02:20 | |
*** Quits: Paul_46 (~IceChat09@cpc1-enfi15-2-0-cust580.hari.cable.virginmedia.com) (Ping timeout: 246 seconds) | 03:06 | |
*** Quits: giallu (~giallu@fedora/giallu) (Read error: Operation timed out) | 03:07 | |
*** Joins: Paul_46 (~IceChat09@cpc1-enfi15-2-0-cust580.hari.cable.virginmedia.com) | 03:43 | |
*** Joins: giallu (~giallu@fedora/giallu) | 07:17 | |
*** Quits: giallu (~giallu@fedora/giallu) (Ping timeout: 276 seconds) | 10:21 | |
*** Joins: BD- (BD-@78-22-91-42.access.telenet.be) | 11:00 | |
*** Quits: BD- (BD-@78-22-91-42.access.telenet.be) () | 11:22 | |
*** Joins: BD- (BD-@78-22-91-42.access.telenet.be) | 11:25 | |
*** Quits: BD- (BD-@78-22-91-42.access.telenet.be) (Client Quit) | 11:27 | |
*** Joins: BD- (BD-@78-22-91-42.access.telenet.be) | 11:27 | |
*** Quits: BD- (BD-@78-22-91-42.access.telenet.be) () | 11:56 | |
*** Joins: BD- (BD-@78-22-91-42.access.telenet.be) | 11:56 | |
*** Quits: BD- (BD-@78-22-91-42.access.telenet.be) (Client Quit) | 11:59 | |
*** Joins: BD- (BD-@78-22-91-42.access.telenet.be) | 12:01 | |
*** Joins: giallu (~giallu@fedora/giallu) | 13:19 | |
*** Joins: dregad (b2c5e92b@gateway/web/freenode/ip.178.197.233.43) | 14:19 | |
*** Quits: BD- (BD-@78-22-91-42.access.telenet.be) () | 14:41 | |
dregad | jreese - you there ? | 14:42 |
---|---|---|
jreese | yessir | 14:42 |
dregad | hey man | 14:43 |
dregad | how are things ? | 14:43 |
jreese | pretty good, how are you? | 14:43 |
dregad | can't complain - on vacation :) | 14:43 |
dregad | spent the day at Europa-Park (kinda disneyland) with the family | 14:44 |
dregad | they're sleeping now :-) | 14:44 |
dregad | i've been getting daily alerts from linode regarding disk io rate lately | 14:45 |
jreese | yeah, I think they're being triggered by the daily backups | 14:46 |
dregad | was looking into it, and looking at graphs, it seems that things have been getting worse (IO steadily increasing over past 30 days, see https://manager.linode.com/linodes/graphs/mantisbt_org?date=monthly | 14:47 |
dregad | you're right they are triggered by the backup | 14:47 |
dregad | but on a "normal" load (i.e. looking at monthly graphs for earlier in the year), the backups did not trigger any alert | 14:48 |
dregad | if it were memory i would say there's a memory leak somewhere, but since it's IO, it must be a file getting bigger and bigger | 14:48 |
dregad | I don't think the DB's been increasing in size, so it's not the backups themselves using more bandwidth | 14:49 |
dregad | the activity on the tracker has been quite low, and the forums not more busy than usual either | 14:50 |
jreese | might be log files | 14:50 |
jreese | or database sizes growing | 14:50 |
dregad | I already ruled out the latter | 14:51 |
jreese | Ok | 14:51 |
dregad | log files make sense, but in /var/log I don't see anything unusual | 14:51 |
Paul_46 | dregad: lo | 14:51 |
jreese | well, /var/log/apache/ and mail.log* will be continuously growing with traffic to the site and tracker | 14:52 |
Paul_46 | we do a daily backup ? :PPP | 14:52 |
* dregad MOOs Paul_46 | 14:52 | |
jreese | not sure if the size is what triggers the extra IO during a backup | 14:52 |
* dregad kicks Paul_46 | 14:52 | |
jreese | Paul_46: Victor enabled the Linode-native system backups | 14:52 |
dregad | I was wondering if it could be the IRC logger | 14:53 |
dregad | Looking at http://www.mantisbt.org/irclogs/ there is a HUGE log file 'cron.log' | 14:53 |
dregad | nearly 300 megs atm | 14:54 |
Paul_46 | downloading it | 14:54 |
Paul_46 | shows lots of 'cp' statements | 14:54 |
Paul_46 | :) | 14:54 |
dregad | yep | 14:54 |
Paul_46 | bw is free thoiugh right? | 14:54 |
dregad | dunno | 14:54 |
dregad | think there's a quota | 14:55 |
dregad | jreese as I'm not too familiar with how you've setup the IRC logs (if I remember correctly what you told me before, your own server is pushing the files to mantisbt.org, right ?) | 14:57 |
dregad | maybe this log file should not be copied over, or if yes, then maybe it could be truncated (or rotated) | 14:57 |
dregad | wdyt? | 14:58 |
Paul_46 | could just copy the logs from my server | 14:58 |
Paul_46 | ;p | 14:58 |
Paul_46 | hmm, that files still downloading | 14:59 |
dregad | lol@Paul_46 | 15:00 |
jreese | dregad: yeah, that file should probably get truncated or ignored or something, didn't realize it was growing that much | 15:00 |
dregad | Paul_46 - actually I think the right way to do things would be to setup a bot directly on mantisbt.org | 15:01 |
Paul_46 | I think we should stop advertising gambling sites on mantisbt.org | 15:03 |
* dregad couldn't agree more | 15:03 | |
Paul_46 | haha | 15:04 |
Paul_46 | that's funny | 15:04 |
Paul_46 | downloading that cron.log crashed IE | 15:04 |
Paul_46 | so it's started again | 15:04 |
dregad | serves you right for using that POS browser | 15:04 |
Paul_46 | that's like 500MB of download | 15:04 |
dregad | don't bother downloading it | 15:04 |
Paul_46 | or something | 15:04 |
Paul_46 | I wanna know if it crashes again | 15:04 |
Paul_46 | i'll attach a debugger and check it's not av plugin | 15:05 |
dregad | you like beating on a dead horse don't you | 15:05 |
Paul_46 | i code on mantis ;p | 15:05 |
dregad | lol | 15:05 |
jreese | zing | 15:05 |
dregad | jreese considering that the logs are being refreshed every 10 minutes iirc, that's 300MB * 6 per hour * 24 = 42 GB... | 15:06 |
dregad | what do you mean 'zing', you zapped the file ? | 15:06 |
Paul_46 | http://www.fosspatents.com/2012/07/apple-seeks-25-billion-in-damages-from.html | 15:07 |
jreese | no, "zing" as in "good insult" | 15:07 |
Paul_46 | apple want 2.5b$ from samsung | 15:07 |
dregad | so what else is new, these 2 have been going at each other for months now | 15:08 |
jreese | dregad: considering that it uses rsync, it's not quite 42GB ;) | 15:08 |
dregad | well i don't know what it uses, so :P | 15:09 |
jreese | the script is synced alongside the results :P | 15:09 |
* dregad should learn Python some day | 15:10 | |
dregad | and | 15:11 |
jreese | it's not that difficult :P | 15:11 |
* dregad should also have checked the shell script file next to the python file too ;) | 15:11 | |
*** Parts: orlok (~orlock@182.160.128.150) () | 15:14 | |
dregad | anyway, jreese could you try truncating that log, to see if the IO goes down | 15:14 |
jreese | sure | 15:14 |
dregad | cheers | 15:14 |
Paul_46 | anyway, it's kinda amusing no one's ever given me details since we moved server | 15:15 |
Paul_46 | been over a year now | 15:15 |
jreese | you haven't asked | 15:15 |
jreese | dregad: crontab updated to truncate the log on every run, and also only output stderr, rather than stderr + stdout | 15:19 |
dregad | thanks john | 15:20 |
jreese | np | 15:20 |
dregad | we'll see if I was right or not at the next backup I guess | 15:20 |
Paul_46 | btw | 15:21 |
Paul_46 | not sure what to do about the dude with he's security issues | 15:21 |
Paul_46 | Globus Healthy & Safe Medifast Diet We like this mommy blog a lot. | 15:21 |
Paul_46 | and wtf | 15:21 |
jreese | we were planning to do something? :P | 15:21 |
* Paul_46 shrugs | 15:21 | |
dregad | which dude are you talking about here | 15:22 |
Paul_46 | anyway | 15:24 |
Paul_46 | i was saying last night or well looking last night | 15:24 |
Paul_46 | and trying to explain to bd- where things were | 15:24 |
Paul_46 | he asked about why we have a next, master, master-1.2 and some other stuff | 15:25 |
dregad | LOL | 15:25 |
Paul_46 | so I through in the master-2.x and tried to do some history | 15:25 |
dregad | I don't think anybody knows ;) | 15:25 |
jreese | dregad: he sent a laundry list of vulnerabilities that were either issues with our upstream libraries, or unintelligible mess of complaints | 15:25 |
Paul_46 | yea | 15:25 |
Paul_46 | he's refined it somewhat | 15:25 |
dregad | ah yes, that | 15:25 |
Paul_46 | but still semi similar | 15:25 |
dregad | I remember now | 15:25 |
Paul_46 | but yea | 15:27 |
Paul_46 | not quite sure what to do with issues in upstream lib's we getting rid of | 15:27 |
Paul_46 | at a point when it's a mess | 15:28 |
Paul_46 | dregad: i looked last night at github's stats | 15:28 |
Paul_46 | whilst talking to bd | 15:28 |
*** Quits: giallu (~giallu@fedora/giallu) (Ping timeout: 276 seconds) | 15:28 | |
dregad | what about nothing | 15:28 |
Paul_46 | specifically: https://github.com/mantisbt/mantisbt/graphs/contributors | 15:28 |
Paul_46 | if I go back to Jan 2009 to now | 15:29 |
Paul_46 | active commiters: dhx, me, jreese, you, rombert, siebrand, victor [ based on >100 commits in 3 years] | 15:29 |
Paul_46 | if I come forward a year and use 60 i.e. 30 a year | 15:30 |
dregad | yes, and if you look at past 6 months, it's 95% me and rombert | 15:30 |
Paul_46 | active commiters: you, dhx, rombert, siebrand, jreese | 15:30 |
dregad | and 99.999% of what rombert does is soap ;) | 15:30 |
Paul_46 | siebrand can be excluded | 15:30 |
Paul_46 | jreese can be excluded | 15:30 |
dregad | jreese is mostly the releases | 15:30 |
dregad | dhx is a few security patches mostly | 15:31 |
dregad | and 'next' | 15:31 |
Paul_46 | so since Jan 2010, you have 181 commits, rombert is 117 | 15:31 |
Paul_46 | (dhx is 146) | 15:31 |
Paul_46 | interestingly, if I do from jan 2010 on my branch idhx has done 268 commits [not sure how that works], and i've done 211 | 15:33 |
dregad | I never actually looked at these graphs, they're nice... can you filter the contributors by date ? | 15:33 |
Paul_46 | yea | 15:34 |
Paul_46 | if you drag on graph date range | 15:34 |
jreese | dregad: the graphs have had significant improvements in the last few months; they weren't even half this good before then | 15:35 |
Paul_46 | yep | 15:35 |
Paul_46 | now i'm unsure where i stand on mantis atm | 15:36 |
Paul_46 | as I kinda enjoy it | 15:36 |
Paul_46 | but as far as i'm concerned, 1.2 and 1.3 are dead | 15:36 |
dregad | really cool these graphs, like them | 15:36 |
dregad | these github dudes are good | 15:36 |
jreese | yep | 15:37 |
dregad | Paul_46 - the problem, as we have discussed several times in the past, is the lack of vision on 'what's next' (no pun on branch naming intended) | 15:37 |
jreese | dregad: that's what I said yesterday :P | 15:37 |
dregad | you do your thing, dhx does his | 15:38 |
dregad | and I maintain 1.2.x because I use it and that's where issues arise | 15:38 |
jreese | "<jreese> [15:23] one thing we've always lacked was a clear set of goals for a release; eg, either a hard deadline or a hard set of features to release" | 15:38 |
Paul_46 | yep | 15:38 |
Paul_46 | dregad: yes, but we've had new features go into 1.2 | 15:40 |
* dregad pleads guilty | 15:41 | |
dregad | but your Honor, it's because 1.3/2.x/next is so far away - the People *need* features | 15:41 |
Paul_46 | so there's new features in 1.2 that personally I think we should deprecate for 1.3 and drop | 15:41 |
Paul_46 | (in some cases) | 15:41 |
Paul_46 | or at least change how they work | 15:41 |
Paul_46 | and that lack of a plan/vision isn't helping | 15:42 |
Paul_46 | hey, one of the things i'd like to add is a free open source mobile interface | 15:42 |
Paul_46 | apart from whether that would kill off mantistouch or not | 15:43 |
Paul_46 | i'd also add a different mobile api for that to what got added between 1.2.8 and 1.2.9 or whatever it was | 15:43 |
Paul_46 | the *easy* way for dealing with that would be to set up mantisbt2.org | 15:43 |
Paul_46 | to differeniate between the two versions of the api's | 15:44 |
Paul_46 | or not bother | 15:45 |
jreese | user confusion ftw | 15:46 |
Paul_46 | jreese: well given that we'd pretty much agreed for mantis 2.0 we'd probably want to drop the current soap api | 15:47 |
Paul_46 | might be less confusing then releasing a 1.3 | 15:47 |
dregad | I don't think it's a good idea to "officially" fork by setting up a separate site, etc | 15:48 |
dregad | it would not only even further dilute and shrink the team, but also as jreese mention, cause user confusion and likely very few would follow to the new version as it essentially breaks the existing | 15:49 |
Paul_46 | well, we have that problem anyway | 15:50 |
Paul_46 | as 1.3/2.0/next [whichever you want to call it] is likely to change db api in a way that woudl require some level of plugin updates | 15:50 |
Paul_46 | and i'd like to look at whether it's possible to tidy up internal structure to make it easier to do things like mobile versions and soap api's | 15:51 |
Paul_46 | and xmlrpc and json etc if people want | 15:51 |
Paul_46 | dregad: and in terms of that, lets clarify a bit | 15:54 |
Paul_46 | November 2009 we had a 'web server api v2 brainstorming' thing on google wave | 15:54 |
Paul_46 | and it listed some goals | 15:54 |
Paul_46 | Goal Number 1: Authentication - don't use the users password | 15:55 |
Paul_46 | Goal 2: Authorization - ability to limit access to an application | 15:55 |
Paul_46 | Goal 3: performance and scability - efficient memory use, throttling, batching | 15:56 |
Paul_46 | Goal 4: Versioning - server side api version, clients provide schema they support, server reponses in version expected by client | 15:56 |
Paul_46 | list went on with another 7-8 things | 15:57 |
Paul_46 | now personally, I like goal 1, and i think it's important | 15:57 |
* dregad wasn't around back then | 15:57 | |
Paul_46 | user should log into mantis | 15:57 |
Paul_46 | generate an "api key(s)" | 15:57 |
jreese | dregad: yeah, I think that was also before dhx | 15:57 |
Paul_46 | webservice should then use that key | 15:57 |
Paul_46 | jreese: dont think so | 15:58 |
Paul_46 | he's in the discussion actually | 15:58 |
jreese | oh, ok | 15:58 |
Paul_46 | anyway, given the number of password leaks etc | 15:58 |
Paul_46 | would you agree goal 1 is probably quite important and reasonable a goal? :) | 15:58 |
Paul_46 | if no, i'd be interested to hear your argument on that one | 15:58 |
jreese | Paul_46: it might be easier to just implement an OAuth provider :P | 15:59 |
Paul_46 | if yes, I suspect implementing that goal would break the current soap api | 15:59 |
Paul_46 | as auth method would probably change from user,pass to authkey if nothing else | 15:59 |
jreese | now that there are some good standards around public API authentication and authorization | 15:59 |
Paul_46 | nod | 15:59 |
* dregad agrees | 16:00 | |
Paul_46 | on the mantis-2 stuff i'd done, i'd basically gone rm -rf soap | 16:00 |
Paul_46 | not because i necessarily want it gone | 16:00 |
Paul_46 | but more, I want to be able to think about mantis and not worry about stuff I can't easily test | 16:00 |
dregad | quite frankly Paul, I think the only problem with your branch is that it has diverged too much from current 1.2.x | 16:01 |
Paul_46 | and work on seeing if it's possible to modernise the legacy core | 16:01 |
dregad | mostly due to my work, it's true | 16:01 |
Paul_46 | but that's also kinda the aim | 16:01 |
Paul_46 | there's a lot in 1.2 that was done for a reason | 16:01 |
dregad | the advantage of 'next' is that dhx has frequently rebased it on top of master | 16:01 |
Paul_46 | in some cases a good reason | 16:01 |
Paul_46 | i'm pretty sure that's got to the point where even dhx isn't keen on the file system layout in it though | 16:02 |
dregad | anyway, switching to your branch would cause either | 16:02 |
dregad | a) loss of hundreds of commits you have not ported | 16:02 |
Paul_46 | i'm not overly concerned about that | 16:02 |
Paul_46 | as, from our stats earlier | 16:02 |
dregad | or b) generate a lot of work to recode stuff | 16:02 |
Paul_46 | 1:30.09] <Paul_46> if I come forward a year and use 60 i.e. 30 a year | 16:03 |
Paul_46 | [11:30.12] <dregad> yes, and if you look at past 6 months, it's 95% me and rombert | 16:03 |
Paul_46 | [11:30.29] <Paul_46> active commiters: you, dhx, rombert, siebrand, jreese | 16:03 |
Paul_46 | [11:30.37] <dregad> and 99.999% of what rombert does is soap ;) | 16:03 |
Paul_46 | so twisting those numbers a little | 16:03 |
Paul_46 | 95% of the stuff in mantis is you+rombert and 100% of rombert's is soap [which needs a rewrite] | 16:04 |
Paul_46 | so if we assuume that you and rombert are roughly 50-50 | 16:04 |
Paul_46 | well probably 60-40 | 16:04 |
dregad | https://github.com/mantisbt/mantisbt/graphs/contributors?from=2012-01-01&to=2012-07-22&type=c | 16:04 |
Paul_46 | that's basically 35% of stuff in last 6 months that can be ignored | 16:04 |
dregad | more like 80%+ me | 16:04 |
Paul_46 | and your stuff is a mix of stuff that is minor | 16:06 |
Paul_46 | and stuff that would require porting | 16:06 |
Paul_46 | some of which is already ported, some is no longer relevant | 16:06 |
Paul_46 | in any case, it's 100 commits | 16:06 |
* dregad finished reading yesterday's IRC log transcript - nice to know the story behind the nick :P~~ | 16:06 | |
Paul_46 | heh | 16:07 |
Paul_46 | so 100 commits to 'review' I dont see as a major issue | 16:07 |
dregad | well actually you should go back to the date of your fork | 16:07 |
dregad | I just picked 01-jan-12 | 16:07 |
Paul_46 | yea | 16:07 |
Paul_46 | but in any case | 16:07 |
dregad | In my opinion, you should work with dhx to merge your stuff with 'next' | 16:08 |
dregad | then we can kill 1.3 | 16:08 |
dregad | announce 2.0 | 16:08 |
dregad | and be done with i | 16:08 |
dregad | it | 16:08 |
Paul_46 | i've not seen dhx for a while | 16:08 |
dregad | me neither, but I have not been online much lately | 16:09 |
Paul_46 | and I dont like the current filesystem layout of 'next' | 16:09 |
dregad | bah | 16:09 |
dregad | you'll get used to it | 16:09 |
dregad | just like I'll have to get used to the new db syntax, etc | 16:09 |
Paul_46 | april 2011 | 16:10 |
Paul_46 | 2011-04-11 | 16:10 |
dregad | well that's way over 250 commits | 16:11 |
Paul_46 | i dont see that as an issue though | 16:11 |
dregad | not counting rombert's btw | 16:11 |
Paul_46 | as a bunch are soap, translation | 16:11 |
Paul_46 | and some ported in already | 16:12 |
Paul_46 | in any case, 250 commits isn't that many | 16:12 |
dregad | still, next has most of them already | 16:12 |
Paul_46 | yea, but there's no way you'd be merging this stuff in | 16:12 |
Paul_46 | easily | 16:12 |
dregad | so merging your branch with next might be less effort than reviewing / porting 250 commits | 16:13 |
dregad | even if manual merge | 16:13 |
dregad | listen i gotta go - it's getting late | 16:14 |
dregad | we should continue this conversation sometime | 16:14 |
dregad | preferably including dhx too | 16:14 |
dregad | @jreese - looks like truncating the log did the trick... https://manager.linode.com/linodes/graphs/mantisbt_org?date=daily | 16:15 |
Paul_46 | 225 commits in this branch | 16:15 |
dregad | how much of it already included in 'next' | 16:15 |
jreese | dregad: excellent | 16:15 |
Paul_46 | dregad: not sure | 16:16 |
Paul_46 | in any case, it's gonna be a pain either way | 16:16 |
Paul_46 | to sort out | 16:16 |
dregad | for sure | 16:16 |
dregad | but if we go 2.0 --> next, we keep the exceptions stuff, which I think is great | 16:17 |
Paul_46 | we're doing exceptions either way | 16:17 |
Paul_46 | the difference is i've not implemented handlers for exceptions | 16:17 |
dregad | whereas if we hand-pick commits from 1.3 to your 2.0, we have to also redo the exceptions | 16:17 |
Paul_46 | ? | 16:17 |
dregad | I mean, re-code what dhx has done in next | 16:18 |
dregad | into your branch | 16:18 |
Paul_46 | and i'm currently throwing exceptions as: | 16:18 |
Paul_46 | throw new MantisBT\Exception\Category_Duplicate(); | 16:18 |
jreese | gad I hate PHP's namespacing choice | 16:18 |
Paul_46 | whereas dhx is doing: | 16:19 |
Paul_46 | use MantisBT\Exception\Security\CSPRNGNotAvailable; | 16:19 |
Paul_46 | and then throw new MasterSaltInvalid(); | 16:19 |
dregad | in other words, your saying you don't like dhx's way ;) | 16:19 |
Paul_46 | personally, if it's possible to do throw new MantisBTException\Duplicate | 16:19 |
Paul_46 | nah | 16:19 |
Paul_46 | not that actually | 16:19 |
Paul_46 | i'd say they are fairly similar | 16:20 |
Paul_46 | i'm not sure what people commonly do | 16:20 |
Paul_46 | additionally i'd not bothered to make exception classes for each exception | 16:20 |
Paul_46 | on the basis I didn't see the point of doing that until we worked out what to throw | 16:20 |
dregad | look i really have to go - it's 1.30 am and the baby's gonna give me a wake-up call in about 5 hrs :-/ | 16:21 |
Paul_46 | and as a minor statement to back that up: | 16:21 |
Paul_46 | https://github.com/mantisbt/mantisbt/blob/next/application/MantisBT/Exception/Filter/FilterTooOldToUpgrade.php | 16:21 |
Paul_46 | $this->responseCode = 500; | 16:21 |
Paul_46 | personally, given exception could get thrown by a cmdline cron job | 16:21 |
Paul_46 | i'm not sure it makes sense to set a 'responseCode' in the exception | 16:21 |
Paul_46 | $this->type = MANTIS_FATAL; sure | 16:22 |
Paul_46 | or something | 16:22 |
Paul_46 | nn ;) | 16:22 |
dregad | good night | 16:22 |
*** Quits: dregad (b2c5e92b@gateway/web/freenode/ip.178.197.233.43) (Quit: Page closed) | 16:22 | |
*** Quits: Paul_46 (~IceChat09@cpc1-enfi15-2-0-cust580.hari.cable.virginmedia.com) (Quit: ASCII a stupid question, get a stupid ANSI!) | 16:50 | |
*** Quits: sdfjkljkdfsljkl (~sdfjkljkd@static.96.23.63.178.clients.your-server.de) (Remote host closed the connection) | 17:00 | |
*** Joins: sdfjkljkdfsljkl (~sdfjkljkd@static.96.23.63.178.clients.your-server.de) | 17:00 | |
*** Joins: giallu (~giallu@fedora/giallu) | 23:47 |
Generated by irclog2html.py 2.10.0 by Marius Gedminas - find it at mg.pov.lt!