Alfresco discussion and collaboration. Stick around a few hours after asking a question.
Official support for Enterprise subscribers: support.alfresco.com.
Join in the conversation by getting an IRC client and connecting to #alfresco at Freenode. Our you can use the IRC web chat.
More information about the channel is in the wiki.
More help is available in this list of resources.
2017-03-03 08:15:12 GMT <yreg> Good morning everyone !
2017-03-03 08:15:26 GMT *** yreg is now known as Guest40133
2017-03-03 10:29:01 GMT <Guest40133> guys what does cache.node.aspectsSharedCache.timeToLiveSeconds=0 means ?
2017-03-03 10:29:12 GMT <Guest40133> cache does not live ?
2017-03-03 10:29:17 GMT <Guest40133> cache does not expire ?
2017-03-03 10:32:21 GMT <Guest40133> Also, is there a significant performance cost diff for identifying elements to be evicted from cache between LFU and LRU policies ?
2017-03-03 10:51:27 GMT <mrks_js> bhagyas: u up? :)
2017-03-03 10:52:06 GMT <mrks_js> bhagyas: i thought you guys developed a cad file previewer plugin? but i can't find anything on your page. only the zoomable plugin
2017-03-03 10:54:16 GMT <bhagyas> mrks_js: Hi
2017-03-03 10:54:17 GMT <bhagyas> yeah
2017-03-03 10:54:24 GMT <bhagyas> It's there on the GitHub
2017-03-03 10:54:36 GMT <bhagyas> called Alfresco Vector Transformations Module
2017-03-03 10:54:52 GMT <bhagyas> Feel free to add custom transformers there
2017-03-03 10:54:53 GMT <bhagyas> :)
2017-03-03 10:55:06 GMT <mrks_js> ah, zoomable has the vector transform
2017-03-03 10:55:09 GMT <mrks_js> cool, thanks : )
2017-03-03 10:55:16 GMT <bhagyas> you have to use Zoomable for the previewing
2017-03-03 10:55:25 GMT <bhagyas> since its independent of the file type
2017-03-03 10:55:29 GMT <mrks_js> i see!
2017-03-03 10:55:35 GMT <bhagyas> :)
2017-03-03 10:57:13 GMT * Guest40133 is attentively waiting for someone like AFaust to pop up and provide some insight <grin> !
2017-03-03 10:57:45 GMT <AFaust> Good idea to mention me - then the task bar icon flashes so I actually do notice
2017-03-03 10:58:13 GMT <fwu> hello all!
2017-03-03 10:58:35 GMT <AFaust> timeToLiveSeconds=0 means that elements will not be expired after a certain time
2017-03-03 10:59:13 GMT <Guest40133> Would it be "dangerous" to set it on nodes and properties caches as well ?
2017-03-03 10:59:26 GMT <AFaust> In that case cache entries are only removed when it is prudent from a logical perspective (e.g. data modified/removed from DB) or the cache is full and evictions occur
2017-03-03 11:00:37 GMT <AFaust> No, it definitely is not dangerous to set it on nodes and properties, and I would actually recommend it. I have been working with a customer in the last couple of months doing some intensive performance testing, and the default timeToLiveSeconds on those caches really make the caches loose their effectiveness
2017-03-03 11:00:43 GMT <Guest40133> Ah! I can see it is already set on properties transactional caches
2017-03-03 11:01:10 GMT <Guest40133> Thought so as well !
2017-03-03 11:01:44 GMT <AFaust> I have not done any tests with LFU as opposed to LRU. Never considered LFU a sensible alternative
2017-03-03 11:02:32 GMT <Guest40133> I think that would be related to the way the system is actually used
2017-03-03 11:03:56 GMT <Guest40133> one last question what is childByNameSharedCache
2017-03-03 11:04:22 GMT <Guest40133> and what does explain why I have only 5.6% hit rate ?
2017-03-03 11:08:57 GMT <AFaust> childByNameSharedCache is a simple lookupCache for navigating from a node to a specific child node using that child node name
2017-03-03 11:10:07 GMT <AFaust> The hit rate on that cache will be very low e.g. if you mostly access different children or from different parents...
2017-03-03 11:12:13 GMT <AFaust> The primary use case of that cache is to optimise NodeService.getChildByName()/deleteNode()/deleteStore() operations
2017-03-03 11:15:19 GMT <Guest40133> Hmm, do you think it is worth making that particular cache bigger then? on my system it has a limit of 130,000
2017-03-03 11:15:36 GMT <Guest40133> it got called 1.5 M times
2017-03-03 11:15:41 GMT <AFaust> The question is: Is that cache close to the limit? Are there many evictions?
2017-03-03 11:15:58 GMT <Guest40133> and only 81k ware successful hits
2017-03-03 11:16:29 GMT <Guest40133> 329k evictions
2017-03-03 11:16:32 GMT <AFaust> The number of calls is not necessarily relevant because those calls are aggregated from the start of the system and never reset - so e.g. a system running for months will naturally have a very high number of calls
2017-03-03 11:16:49 GMT <Guest40133> the systems is running for a week or so
2017-03-03 11:18:06 GMT <AFaust> For this cache in particular, the number of calls does not provide a sensible indication of relevance, because the cache will always be called first before the DB is queries, and most calls will be for child nodes that do not exist at all
2017-03-03 11:19:09 GMT <AFaust> So since you have 329k evictions and a limit of 130k, that means from the 1.5M calls only (459k + hit count) actually matched existing child nodes
2017-03-03 11:19:27 GMT <Guest40133> Ok, judging by mmetrics for other caches, 10.8M evictions for nodes shared cache for example, tuning the childByNameSharedCache won't be as effective (in terms of performance) as tuning the ones with very high eviction rates
2017-03-03 11:21:34 GMT <Guest40133> humm good catch! but some entries could have been evicted multiple times
2017-03-03 11:21:42 GMT <AFaust> Especially since the nodes shared cache is essentially the "key" cache in Alfresco.
2017-03-03 11:21:44 GMT <Guest40133> but still a very valid point !
2017-03-03 11:22:15 GMT <AFaust> True, there could be some overlap but the estimate of an upper bound is still good to know
2017-03-03 11:23:06 GMT <Guest40133> agreed
2017-03-03 11:24:21 GMT <Guest40133> do you, by any chance, know of a way to optimize ACL checks, I have the feeling that those are also a key factor to my performance issues
2017-03-03 11:24:52 GMT <Guest40133> I can see some ACL caches
2017-03-03 11:25:02 GMT <Guest40133> with very low limit
2017-03-03 11:25:50 GMT <Guest40133> and no information about hits/misses nor evictions
2017-03-03 11:26:55 GMT <AFaust> The standard ACL checks are actually quite efficient from my profiling experience. It often are customer/3rd-party customisations that wreck ACL checking performance, and a lot of the time people also mistake the DB load time to pre-cache nodes with the ACL check time...
2017-03-03 11:27:27 GMT <AFaust> So if you "have the feeling" that ACL checks are slow you should first try to verify that feeling by doing some rudimentary CPU / execution time sampling
2017-03-03 11:27:51 GMT <AFaust> Guest40133: Are you using Enterprise Edition?
2017-03-03 11:28:02 GMT <Guest40133> Yes, in a two nodes cluster
2017-03-03 11:28:43 GMT <AFaust> That explains why you don't see statistics for ACL caches - some of the cache implementations used in Enterprise "suck" in that they give inconsistent / incomprehensible statistic values
2017-03-03 11:28:46 GMT <Guest40133> and I have found that for some requests 18-26 % of CPU time is reserved for permission checks
2017-03-03 11:29:00 GMT <AFaust> For that reason I had to remove those cache details from view in the caches tool.
2017-03-03 11:29:19 GMT <Guest40133> hmm, that explains it then !
2017-03-03 11:29:36 GMT <Guest40133> I am using OOTBee Support tools by the way
2017-03-03 11:30:01 GMT <AFaust> You need to drill down in the permission check times because there are code paths there that may do bulk fetching of nodes / pre-loading from DB that may inflate permission check times
2017-03-03 11:30:28 GMT <AFaust> I know - OOTBee Support Tools is about the only tool that really allows you to check cache metrics
2017-03-03 11:31:27 GMT <Guest40133> AFaust: FYI in my system, according to solr, I do have twice as many ACL as nodes
2017-03-03 11:32:03 GMT <AFaust> Ehm - ok, how'd you manage to achieve that?
2017-03-03 11:32:26 GMT <Guest40133> and I feel like the the out of the box sizes for ACL caches does not meet my setup
2017-03-03 11:32:29 GMT <AFaust> Technically that would mean each node has its own ACL
2017-03-03 11:32:53 GMT <Guest40133> AFaust: unfortunately it is pretty much like it !
2017-03-03 11:33:10 GMT <AFaust> And due to the way Alfresco handles ACL inheritance the effective number of ACL is always close to 2-times the number of unique ACLs
2017-03-03 11:34:35 GMT <AFaust> But regarding cache sizes: The questions again are "How many items are in the cache?", "How many evictions occur?"...
2017-03-03 11:35:43 GMT <Guest40133> AFaust: this project started back in 2010 I think, and they used a "unconventional" way of setting ACLs: a multi valued property on ther root doctype for list of group who have access to a node, a property propagation startegy and a policy that take that property, disable permission inheritance and set permissions according to that property's content
2017-03-03 11:35:45 GMT <AFaust> If you really do have unique ACL per node then there can indeed be a significant overhead in permission checking, as ACLs are not loaded via batches and you'll have a lot of DB roundtrips if the caches are too small to hold on to the data
2017-03-03 11:36:22 GMT <AFaust> Ok - that sounds like the approach that would be necessary to achieve such a state..
2017-03-03 11:37:49 GMT <Guest40133> the existing (shown) metrics in OOTBee ST for ACLs does not seem reliable
2017-03-03 11:38:06 GMT <Guest40133> 53/32 gets ?!
2017-03-03 11:38:45 GMT <Guest40133> lunch time, be back in one hour !
2017-03-03 11:42:40 GMT <AFaust> Ah right - completely blocked out the fact that we don't show statistics for that after I myself explained why... got distracted by issues I am having with a customers docker setup
2017-03-03 12:17:45 GMT <Guest40133> AFaust: I can help with docker if you want
2017-03-03 12:18:17 GMT <Guest40133> I can understand that you do not show metrics that you think are inaccurate
2017-03-03 12:18:58 GMT <Guest40133> but are the metrics shown accurate ?
2017-03-03 12:19:25 GMT <AFaust> Doubt it since it is very likely instance/setup specific. For some reason a tomcat in one container cannot establish a connection to a Postgres in another.
2017-03-03 12:19:56 GMT <AFaust> Though from a network perspective nothing is being blocked and the port is exposed (I can connect from a different host)
2017-03-03 12:20:43 GMT <Guest40133> AFaust: does these containers belong to a common network ?
2017-03-03 12:20:54 GMT <AFaust> Regarding the metrics accuracy: It is hard to verify the specific numbers without taking Alfresco apart and measuring yourself for comparison
2017-03-03 12:21:14 GMT <Guest40133> https://docs.docker.com/engine/userguide/networking/work-with-networks/
2017-03-03 12:21:15 GMT <alfbot> Title: Work with network commands - Docker Documentation (at docs.docker.com)
2017-03-03 12:21:36 GMT <AFaust> The metrics we chose to hide were obviously inaccurate, showing sometimes double the amount of gets
2017-03-03 12:21:51 GMT <AFaust> all containers run on the same host
2017-03-03 12:22:31 GMT <Guest40133> AFaust: that does not necessarily mean they are on the same network, unless they are defined on the same docker-compose.yml
2017-03-03 12:22:37 GMT <AFaust> config is managed via docker-compose which was not touched in weeks
2017-03-03 12:23:27 GMT <AFaust> Differet docker-compose.yml for DB and affected host - but also unchanged in weeks as far as I can see
2017-03-03 12:24:00 GMT <AFaust> Worked fine 14days ago
2017-03-03 12:25:14 GMT <AFaust> Already tried accessing the DB via the public interface of the docker host (to rule out issues with networking)
2017-03-03 12:28:55 GMT <AFaust> BTW: Default connection attempt is not by calling DB container directly, instead call is against private IP of host which should then forward to DB container
2017-03-03 12:30:18 GMT <Guest40133> you'd better inspect the automatically created network by those compose files
2017-03-03 12:30:46 GMT <Guest40133> by default, I doubt that containers from one composition can access containers from other compositions
2017-03-03 12:30:58 GMT <Guest40133> without setting that up explicitely
2017-03-03 12:31:44 GMT <Guest40133> and setting that explicitly in docker-compose files will persist, but if done manually via CLI, it does not persist
2017-03-03 12:37:12 GMT <Guest40133> AFaust: FYI, even if you did not change the compose files, docker-compose up results in recreating updated containers, so if any of your passed env variables depends on the host env / any of the images/tags got updated the container gets recreated !
2017-03-03 12:37:33 GMT <Guest40133> and any manual networking setup is lost !
2017-03-03 12:44:38 GMT <AFaust> There is zero manual setup
2017-03-03 12:45:49 GMT <Guest40133> if you get inside alfresco container can you ping the db container
2017-03-03 12:46:23 GMT <Guest40133> docker exec -it compose_alfresco-platform_1 bash
2017-03-03 12:47:09 GMT <Guest40133> is there a chance that the IPs have changed ?
2017-03-03 12:50:58 GMT <AFaust> Container IPs have definitely changed - the network setup has always been setup to not rely on that
2017-03-03 12:51:06 GMT <AFaust> Turns out it is a JDBC issue
2017-03-03 12:53:12 GMT <AFaust> There has been an update of the DB version (though I still see the same image being referenced) and the old JDBC driver cannot connect which is reported like a generic network issue instead of a proper "incompatible version detected"
2017-03-03 12:55:07 GMT <AFaust> ^^ At least as far as I could check / confirm from the docker host - have to verify that from inside the container
2017-03-03 13:11:35 GMT <Guest40133> AFaust: if the image is referenced without a tag, then the tag latest is always used
2017-03-03 13:12:16 GMT <Guest40133> AFaust: if the image is referenced WITH a tag, and that tag changes, then the container is recreated with the latest image
2017-03-03 13:12:38 GMT <Guest40133> that is on a docker-compose up [-d]
2017-03-03 13:13:35 GMT <AFaust> It is reference with a tag and the tag is the same
2017-03-03 13:13:51 GMT <AFaust> Starting to think some of the iptables madness docker does is not working out
2017-03-03 13:14:26 GMT * AFaust is reminded why he avoids docker most of the time...
2017-03-03 13:29:44 GMT * IanW1 is reminded of https://thehftguy.wordpress.com/2016/11/01/docker-in-production-an-history-of-failure/ - in summary don't use docker in production esp. for databases
2017-03-03 13:29:45 GMT <alfbot> Title: Docker in Production: A History of Failure – The HFT Guy (at thehftguy.com)
2017-03-03 13:31:03 GMT <Guest40133> IanW1: we have been using it in production for almost 2 years now, without issues
2017-03-03 13:37:27 GMT <resplin> fwu: I saw your question from yesterday. It sounds like you have a hardware problem.
2017-03-03 13:37:49 GMT <resplin> fwu: In general, you shouldn't force apply amps. That allows non-compatible amps to overwrite each other.
2017-03-03 13:38:08 GMT <resplin> But that IO error sounds like an actual problem writing to disk, and not anything with Alfresco.
2017-03-03 13:41:31 GMT <IanW1> Guest40133: I only use docker in dev/test (and not for Alfresco) so don't have any direct experience/comments to make but the article made for interesting reading
2017-03-03 13:48:29 GMT <AFaust> IanW1: I am aware of that blog post. It was one of the most recent blog posts I could find that comes close to sum up what I feel about docker when someone asked my "Why do you hate it?"
2017-03-03 13:49:50 GMT <AFaust> To be honest - when it works as smooth as Guest40133 points out it has been running for them, it is perfectly fine. I just have issues with the complexities when things go wrong...
2017-03-03 13:50:43 GMT <DarkStar1> Hello everyone
2017-03-03 13:50:45 GMT <IanW1> I only use docker in conjunction with packer so am fairly ambivalent about it
2017-03-03 13:53:04 GMT <Guest40133> fwu: I did not see your original question
2017-03-03 13:53:17 GMT <AFaust> Ah well - now that someone at the customer was able to help me (haven't yet figured out "how" or "what" exactly he did), I can begin with the task I wanted to start 4 hours ago
2017-03-03 13:53:35 GMT <Guest40133> but if you are running on windows server I doubt resplin analysis is right
2017-03-03 13:53:49 GMT <Guest40133> I think it is more likely to be permission issue
2017-03-03 13:54:08 GMT <Guest40133> so try running cmd as admin before applying amps
2017-03-03 13:54:58 GMT <Guest40133> ** before calling the apply amps batch
2017-03-03 13:56:16 GMT <Guest40133> as for forcing apply amps, it is generally not the intended way for applying amps but sometimes necessary for overriding parts of alfresco that do not expose extension points
2017-03-03 13:56:46 GMT <Guest40133> so I would say you should check with the addon release notes/docs/readme/owner first
2017-03-03 13:58:56 GMT <Guest40133> having to force amps is not always a result of sloppy code, it is sometimes needed, and sometimes an unconvenience of having two addon developped separately having an artifact with the same name (can be a 3rd party lib that does not exist in alfresco by default and that both amps include)
2017-03-03 13:59:04 GMT <AFaust> fwu, Guest40133: If an addon overrides parts of Alfresco you should notify the developer - in most cases there are better alternatives without overriding, even if the addon e.g. contains a core code fix for an issue that Alfresco refuses to fix (which seems to be on the rise)
2017-03-03 13:59:14 GMT <resplin> You are right. Could be a permission issue. But it's a problem writing data in the OS. It isn't Alfresco.
2017-03-03 13:59:55 GMT <resplin> AFaust: I wouldn't say it is on the rise. It's just that we are being better at making the decision instead of never getting around to it.
2017-03-03 14:00:06 GMT <resplin> Total number of fixes h as gone up over the past 12 months.
2017-03-03 14:00:24 GMT <resplin> But we do give more bad news.
2017-03-03 14:03:24 GMT <AFaust> So the decision to not fix something that is certifiably not working correct is on the rise. Not saying that having decisions sooner is a bad thing, but we all know "won't fix" are pretty much final in practice and that issue is going to be in the product for years...
2017-03-03 14:06:15 GMT <AFaust> Might have been not as bad if the "won't fix" would be differentiated into "not a real issue - never gonna do it" and "real issue - we just don't care enough right now" (aka "other priorities")
2017-03-03 14:08:28 GMT <AFaust> Otherwise how are you going to go about revisiting "known issues" in a few years? Unless someone re-reports an issue (which is unlikely if they find it already rejected) you'd have to go through all the "won't fix" issues again - which even I would refuse to do, so don't expect to ever be done
2017-03-03 14:10:23 GMT <Guest40133> resplin: AFaust make a very valid points here !
2017-03-03 14:10:33 GMT <Guest40133> point*
2017-03-03 14:10:34 GMT <AFaust> My biggest regret is that the community is too fragmented / individualistic to really come together and start working on a common "Rejected / Improved Fixes" addon project
2017-03-03 14:11:25 GMT <AFaust> "regret" => "disappointment"
2017-03-03 14:11:52 GMT <Guest40133> Well Partners/EE customer wouldn't make any effort towards an unsupported setup
2017-03-03 14:12:16 GMT <Guest40133> and that's already a huge part of the community AFAIK
2017-03-03 14:13:20 GMT <AFaust> I see that differently - partners / EE customers are maybe the biggest part of the user base, but with regards to community, they are the most underrepresented group of all.
2017-03-03 14:14:33 GMT <AFaust> Apart from Bindu, you and mrks_js, I can't think of other partner employees that engage in the community.
2017-03-03 14:16:32 GMT <bhagyas> AFaust: If you think Order of the Bee is all that is left of Alfresco community, you're wrong
2017-03-03 14:16:43 GMT <AFaust> I don't think that
2017-03-03 14:17:08 GMT <bhagyas> AFaust: And I find your omission of Loftux work disappointing
2017-03-03 14:18:17 GMT <AFaust> Where did I omit it where it would have applied?
2017-03-03 14:18:54 GMT <bhagyas> I myself was a part of the community while I was was working in not one, but two enterprise partners
2017-03-03 14:19:14 GMT <AFaust> You might remember I was also working for a partner...
2017-03-03 14:19:41 GMT <AFaust> Same as with Jan and Florian from fme
2017-03-03 14:20:18 GMT <AFaust> I was talking about the current state though...
2017-03-03 14:20:57 GMT <Guest40133> AFaust: I work for a Partner for instance
2017-03-03 14:21:00 GMT <AFaust> And even the handful of people we can include from history would still not cause EE partners / customer to be represented according to their share of the user base...
2017-03-03 14:21:33 GMT <AFaust> Guest40133: I know - I included you by saying "you" => "Apart from Bindu, you and mrks_js..."
2017-03-03 14:23:01 GMT <Guest40133> Missed that
2017-03-03 14:23:34 GMT <Guest40133> any way, I think I remember seing some thread on the hive list around this thing
2017-03-03 14:23:45 GMT <douglascrp> hey guys
2017-03-03 14:23:48 GMT <douglascrp> interesting discussion
2017-03-03 14:24:13 GMT <bhagyas> AFaust: Also, there is a project that fixes what Alfresco won't fix, called LXCommunity
2017-03-03 14:24:20 GMT <bhagyas> and you're welcome to always contribute
2017-03-03 14:24:40 GMT <Guest40133> and there was even some community member that rushed into creating an artifact (that had his company as group ID, if I remeber right, which led to other members not to contribute)
2017-03-03 14:25:18 GMT <AFaust> bhagyas: Maybe you did not understand what I was referring to. Yes, there is LXCommunity, but there is also Angel, ECM4U and a number of other people doing their own version of fixes
2017-03-03 14:25:28 GMT <Guest40133> and after the ownership by OOTBee got defined, there were no further discussion in that thread AFAIK
2017-03-03 14:26:55 GMT <bhagyas> done is better than perfect and the state is fine, it's alfresco's loss that it misses out on contributions from the community
2017-03-03 14:27:27 GMT <AFaust> Guest40133: I think Tichodroma started with a project and there was some discussion around it. But the entire idea suffered from the same problem most ideas suffer from - too few people (if any) are open to a discussion and joining of forces, maybe for fear of becoming in some form dependent / restricted by the collaboration
2017-03-03 14:27:50 GMT <bhagyas> there is one russia based vendor which keeps Alfresco sane from Aikau madness - well maintained with consistent UI and updates
2017-03-03 14:29:35 GMT <douglascrp> DarkStar1, hey dude
2017-03-03 14:29:50 GMT <douglascrp> how about your issue:
2017-03-03 14:29:57 GMT <douglascrp> were you able to fix it
2017-03-03 14:30:08 GMT <Guest40133> resplin: I really hope Alfresco does revise its policy in regards to accepting contributions (fixes)... it will definitely make your life easier to have to validate/merge patches instead of spending valuable time arguing either or not a confirmed bug should be fixed
2017-03-03 14:31:22 GMT <AFaust> I think resplin may address the topic at BeeCon and be available for discussion then...
2017-03-03 14:31:41 GMT <Guest40133> If you (alfresco) can not attribute commit rights based on merits like other opensource projects, you should at least treat contributions with the some seriousness
2017-03-03 14:32:39 GMT <bhagyas> Guest40133: That will happen, but on a parralell universe ;) or maybe we can expect another talk at next years BeeCon
2017-03-03 14:32:47 GMT <bhagyas> xD
2017-03-03 14:33:45 GMT <Guest40133> bhagyas: There is actually a talk dedicated to contributing to alfresco this beecon... So I am hoping for good news in that reguard
2017-03-03 14:34:18 GMT <bhagyas> Guest40133: I've been to at least three conferences where Alfresco made that promise, looking forward to more ;)
2017-03-03 14:35:30 GMT <bhagyas> instead of malice, I now attribute the broken contribution process to incompetency and unwillingness
2017-03-03 14:35:44 GMT <bhagyas> https://en.wikipedia.org/wiki/Hanlon%27s_razor
2017-03-03 14:35:45 GMT <alfbot> Title: Hanlon's razor - Wikipedia (at en.wikipedia.org)
2017-03-03 14:37:01 GMT <DarkStar1> douglascrp: I literally haven't had the time to test
2017-03-03 14:37:12 GMT <Guest40133> bhagyas: you definitely need to be less hostile/offensive, those are really BIG words
2017-03-03 14:37:31 GMT <bhagyas> yup
2017-03-03 14:37:37 GMT <bhagyas> now I simply ignore
2017-03-03 14:37:38 GMT <bhagyas> :)
2017-03-03 14:37:43 GMT <bhagyas> let's enjoy the ride lol
2017-03-03 14:38:05 GMT <douglascrp> DarkStar1, ok, no probem
2017-03-03 14:38:14 GMT <douglascrp> DarkStar1, let me know how it goes
2017-03-03 14:38:31 GMT <Guest40133> I understand that what is considered to be offensive in one culture is not necessarily the same in other cultures, but still I think you make it too difficult to make up a constructive conversation with you in it !
2017-03-03 14:39:17 GMT <bhagyas> Guest40133: compared to the number of conversations that have not had 'me' in it, I'm sure if they had 'any' conversation about it, something would have happened
2017-03-03 14:39:28 GMT <bhagyas> just pointing out the reality here
2017-03-03 14:40:35 GMT <bhagyas> you have now ended up in a point, if they can't find someone to fix it, they'd just close issue reports worrying abut 'organisational metrics', someone always passes the ball
2017-03-03 14:41:03 GMT <bhagyas> I had a good time looking at a well detailed issue report from AFaust being closed yesterday for no reason xD
2017-03-03 14:41:16 GMT <AFaust> Java 8 Lambda?
2017-03-03 14:41:26 GMT <bhagyas> yups
2017-03-03 14:41:53 GMT <bhagyas> If they granted commit rights to AFaust to fix it, it would probably take him 2 hours
2017-03-03 14:41:56 GMT <bhagyas> xD
2017-03-03 14:42:29 GMT <bhagyas> instead of writing a big argument on how java 8 improves things
2017-03-03 14:43:10 GMT <AFaust> To be fair, the issue would still need to be documented. I currently work at a customer where I already hate the one- or two-liner issues without any background info.
2017-03-03 14:43:39 GMT <AFaust> As much as I hate the same kind of question on community.alfresco.com...
2017-03-03 14:45:14 GMT <AFaust> I honestly can live with that issue not being addressed more than I can live with the "full unicode support" one...
2017-03-03 14:53:39 GMT <Guest40133> AFaust: what full unicode support ?
2017-03-03 14:54:40 GMT <AFaust> There are issues with full 4-byte unicode characters in some parts of the Share application (actually, only in the Surf URLEncoder but that is used in quite a few places)
2017-03-03 14:54:56 GMT <AFaust> https://issues.alfresco.com/jira/browse/ALF-21846
2017-03-03 14:55:07 GMT <AFaust> https://issues.alfresco.com/jira/browse/ALF-21847
2017-03-03 14:55:37 GMT <AFaust> The Search team appears to be addressing the issue - for Share it was determined as "won't fix"
2017-03-03 14:58:53 GMT <AFaust> It is tough to determine how big the impact could be on some regions with special alphabets etc. Using an emoji to demonstrate it might not have been the best choice but at the time I was not aware of a meaningful set of characters that might be affected...
2017-03-03 14:59:49 GMT <AFaust> I only recently learned of IICORE, but still do not really understand what the reach of that is.
2017-03-03 15:02:44 GMT <AFaust> Example of affected characters: http://www.i18nguy.com/unicode/supplementary-test.html
2017-03-03 15:02:45 GMT <alfbot> Title: Unicode Supplementary Test Characters (at www.i18nguy.com)
2017-03-03 15:03:32 GMT <fwu> respli, AFaust: the overwrite is not hte problem. Im doing it with a goal. The problem is the error. More stange is that Im able to run the he apply amps alfresco-mmt.jar instructions withou any error!
2017-03-03 15:04:01 GMT <fwu> once I had something similar in a server... I need to run the apply_amps as an administrator
2017-03-03 15:04:04 GMT <AFaust> fwu: The discussion has gone way off-topic from your original problem I realize...
2017-03-03 15:04:41 GMT <fwu> AFaust, the problem is just the error itself, in my own machine when I already did this a lot of times.
2017-03-03 15:05:02 GMT <fwu> and if I run the commands independently with alfresco-mmt.jar it works...
2017-03-03 15:05:05 GMT <fwu> very strange
2017-03-03 15:05:38 GMT <fwu> I reeboted a lot of times and it doesnt solve either
2017-03-03 15:05:40 GMT <AFaust> Using MMT JAR standalone and Maven plugin are technically quite different.
2017-03-03 15:06:10 GMT <fwu> so, basically, I cant run the apply_amps in my dev machine...
2017-03-03 15:07:04 GMT <AFaust> Unfortunately I don't have any input on apply_amps - I don't use it because in systems I manage it does not exist (#installerboycott)
2017-03-03 15:10:09 GMT <Guest40133> fwu: are you trying to apply amps on windows
2017-03-03 15:10:10 GMT <Guest40133> ?
2017-03-03 15:10:19 GMT <Guest40133> AFaust: interesting to know
2017-03-03 15:10:42 GMT <douglascrp> fwu, what is the error message?
2017-03-03 15:11:50 GMT <Loftux> I'm trying to setup replication, and just getting 401 when running the job. First question, is it really dependent on configuring share-urls ? And if so, why do you create a replication service that is dependent on client ui?
2017-03-03 15:12:32 GMT <Guest40133> Loftux: I do not think it should point to share
2017-03-03 15:13:24 GMT <fwu> yes, on windows
2017-03-03 15:13:40 GMT <fwu> it always worked
2017-03-03 15:14:00 GMT <AFaust> Loftux: The Share-tier URL config is only for linking to the other systems UI in Share views
2017-03-03 15:14:01 GMT <Guest40133> fwu run CMD as Administrator (click right -> run as admin)
2017-03-03 15:14:03 GMT <AFaust> AFAIK
2017-03-03 15:14:12 GMT <Guest40133> then invoke apply amps from there
2017-03-03 15:14:47 GMT <Guest40133> fwu: due to some stupid permissions policies on some stupid windows server distributions
2017-03-03 15:15:01 GMT <AFaust> Other than that - the last time I set up replication (2 years ago) I only had to configure Repo URLs for the actual replication endpoints)
2017-03-03 15:15:05 GMT <Guest40133> it does not work unless you run apply amps as an admin
2017-03-03 15:15:10 GMT <fwu> douglascrp, 02020000 An IO error was encountered during deployment of the AMP into the WAR
2017-03-03 15:15:14 GMT <Loftux> AFaust: OK, that explains it, then I can figure out what actually needs to go there. Now to understand why I get the 401
2017-03-03 15:16:02 GMT <Guest40133> Loftux: I experimented with it a few months back, and the docs were more than enough to cover my needs
2017-03-03 15:16:31 GMT <fwu> Guest40133, I alwyas do that
2017-03-03 15:16:37 GMT <fwu> it doesnt work anymore
2017-03-03 15:16:43 GMT <Guest40133> I just did not apreciate much the fact that the replicated structure on the targes system is immutable
2017-03-03 15:17:22 GMT <Guest40133> fwu it is definitely a security issue with the OS
2017-03-03 15:17:43 GMT <fwu> hmm.. the only change here was my user that now belongs to another domain
2017-03-03 15:17:47 GMT <Guest40133> if you are looking at alfresco you are looking at the wron direction
2017-03-03 15:17:53 GMT <fwu> but I beleive Im a local admin
2017-03-03 15:23:01 GMT <Loftux> I see, the user that you configure to connect pretty much must be an admin user, when I change from my replication user to admin it worked. Even if the user I created for the replication task exist on both servers and have the same access to the site with the folder I was trying to replicate.
2017-03-03 15:25:39 GMT <fwu> but i can leave with this as im able to run the alfresco-mmt.jar commands
2017-03-03 15:27:30 GMT <Guest40133> Loftux: I suppose the replication endpoint on targetsystem has <authentication>admin</authentication>
2017-03-03 15:28:55 GMT <fwu> ppl, in the alfrescoStartevent, the start event and the end even both runs AFTER the user clicks on the action button in the screen right? If so, what is the different between them?
2017-03-03 15:28:56 GMT <mrks_js> fwu: maybe the issue is with the clean_tomcat script? afaik this is being run at some point too
2017-03-03 15:29:08 GMT <fwu> hmm
2017-03-03 15:29:16 GMT <fwu> it can be. I may try run it
2017-03-03 15:29:20 GMT <fwu> let me see
2017-03-03 15:32:05 GMT <Loftux> Guest40133: You are correct. How I love having to create System users with full access…
2017-03-03 15:47:19 GMT <Guest40133> ~seen angelborroy
2017-03-03 15:47:19 GMT <alfbot> Guest40133: angelborroy was last seen in #alfresco 1 day, 23 hours, 42 minutes, and 50 seconds ago: <angelborroy> If I can get an stable network, I’ll broadcast live the session
2017-03-03 15:50:33 GMT <douglascrp> fwu, isn't the drive a virtual one?
2017-03-03 15:52:00 GMT <douglascrp> fwu, I remember about that message once
2017-03-03 15:52:17 GMT <douglascrp> I was using the windows subst command to create a virtual drive where I was developing
2017-03-03 15:53:24 GMT <fwu> no, it local
2017-03-03 15:53:26 GMT <fwu> its
2017-03-03 15:54:29 GMT <fwu> mrks_js, clean tomcat is also working
2017-03-03 15:55:05 GMT <fwu> so, the running the individual commands works
2017-03-03 15:55:42 GMT <fwu> but they dont work while running inside a bat file...
2017-03-03 15:55:59 GMT <fwu> i can find any sense here
2017-03-03 15:56:04 GMT <fwu> can´t
2017-03-03 16:16:56 GMT <fwu> ppl, how to pass a value filled by the user in the startevent screen of the workflow?
2017-03-03 16:18:15 GMT <fwu> execution.getVariable deost seems to work
2017-03-03 16:21:19 GMT <Guest40133> fwu you are talking about embedded activiti right
2017-03-03 16:21:20 GMT <Guest40133> ?
2017-03-03 16:21:38 GMT <fwu> yes
2017-03-03 16:21:43 GMT <Guest40133> is this property set in the model on task type
2017-03-03 16:21:47 GMT <Guest40133> or on an aspect ?
2017-03-03 16:22:21 GMT <fwu> let me see
2017-03-03 16:23:17 GMT <fwu> aspect
2017-03-03 16:24:00 GMT <douglascrp> fwu, sorry, but I don't know what can be wrong
2017-03-03 16:24:45 GMT <fwu> douglascrp, thanks anyway :)
2017-03-03 16:35:15 GMT <fwu> Guest40133, but I have the mandatory -aspect applied
2017-03-03 16:39:53 GMT <Guest40133> then yes, that value should be transferred automatically to your execution
2017-03-03 16:45:27 GMT <resplin> Back to the conversation about accepting patches, there are two important prerequisites we need:
2017-03-03 16:45:41 GMT <resplin> * the ability to accept pull requests instead of diff files
2017-03-03 16:45:56 GMT <resplin> * reliable automated tests that are available for people to test their patches
2017-03-03 16:46:21 GMT <resplin> We have made a lot of progress on both fronts over the last year, which I will be able to talk about at BeeCon.
2017-03-03 16:47:56 GMT <resplin> I share the frustration that Alfresco isn't better at accepting contributions, and I have been working to improve the situation.
2017-03-03 16:48:51 GMT <resplin> I find it equally frustrating when people don't acknowledge the progress we do make.
2017-03-03 16:50:29 GMT <fwu> Guest40133, I try to get the values on the start and end event without success
2017-03-03 16:51:24 GMT <Guest40133> resplin: agreed
2017-03-03 16:51:35 GMT <Guest40133> and thank you for your valuable efforts
2017-03-03 16:52:50 GMT <resplin> We will continue the conversation at BeeCon. People shouldn't expect a "look we've solved all the problems". But we will have a "last year we talked about these things, and we got the following done"
2017-03-03 16:52:59 GMT <resplin> "And we still have left these other important things"
2017-03-03 16:53:21 GMT <Guest40133> resplin: and for the record, I am sure there is a lot of enthousiast in the community that are willing, sometimes even trying hard, to help... and they just do not find how they can do so
2017-03-03 16:53:36 GMT <resplin> You know, similar to the last three times we have had that conversation at a conference. Collaboration is an ongoing thing.
2017-03-03 16:53:46 GMT <Guest40133> :D
2017-03-03 16:53:59 GMT <resplin> Guest40133: I recognize that. That's part of the difficulty in explaining the roadblocks. It's like a bad relationship "it isn't you, it's me".
2017-03-03 16:54:12 GMT <resplin> We have to solve our internal process and organizational issues. Those aren't easy.
2017-03-03 16:55:16 GMT <resplin> Your particular offer of helping with the Git history is an excellent example. We really should do that. But migrating dozens of engineers in a very short window introduces a lot of risk.
2017-03-03 16:55:37 GMT <resplin> Guest40133: I think I saw yesterday that this is yreg, correct?
2017-03-03 16:56:00 GMT <resplin> The technical bit is the easy bit.
2017-03-03 16:56:57 GMT <resplin> Similar situations arise with the ALF Triage. People complain so loudly when we close issues. What they don't realize is that we have a backlog of 80 community raised issues that we are actually working on.
2017-03-03 16:57:15 GMT <resplin> That's a lot of engineering not tied directly to a customer.
2017-03-03 16:57:47 GMT <resplin> At the same time, we only fix a few dozen issues per service pack most of which are customer raised.
2017-03-03 16:57:59 GMT <Guest40133> resplin: need to go now, will check the log later
2017-03-03 16:58:02 GMT <resplin> We simply aren't going to get done most of what people ask us to do.
2017-03-03 16:58:14 GMT <resplin> No worries. Have a good evening.
2017-03-03 16:59:00 GMT <AFaust> resplin: I my kind of "software engineering school of thought" that I "grew" up in (e.g. was trained), you *never* close a legitimate issue. You merely re-prioritise it. Otherwise you are sweeping issues under the rug and increase the size of your blindspots...
2017-03-03 17:00:17 GMT <resplin> AFaust: I understand that train of thought. But I don't think it is realistic.
2017-03-03 17:00:20 GMT <AFaust> So forgive me if I started to become one more person that may frustrate you when I "don't acknowledge the progress [that was made]" but subjectively it has gone from bad to worse instead of the other way around.
2017-03-03 17:00:44 GMT <resplin> And you were one of the people that was frustrated with how long issues would stay open.
2017-03-03 17:01:09 GMT <AFaust> To be clear: Open without ANY reaction
2017-03-03 17:01:11 GMT <resplin> One of the quotes I remember from this channel is "if you would be clear what you won't do, then at least we know what we need to work around"
2017-03-03 17:01:25 GMT <resplin> Though that might have been Loftux
2017-03-03 17:01:40 GMT <resplin> Now every issue gets reviewed almost every week.
2017-03-03 17:01:50 GMT <resplin> Though some take a long time to figure out what to do with.
2017-03-03 17:02:03 GMT <resplin> Deciding "no we won't ever do that" helps us to prioritize the issues we will actually do.
2017-03-03 17:02:08 GMT <resplin> I should say "may actually do".
2017-03-03 17:02:57 GMT <AFaust> No, it could have been me as well - I possibly should have made myself a bit more clear back then, but putting something on a backlog with a target version waay in the future would have been the information needed regarding issues "Alfresco won't do" (soon)...
2017-03-03 17:03:34 GMT <resplin> AFaust: The nice thing about talking with you is that we can have a clear discussion even when we disagree. I don't find that offensive.
2017-03-03 17:03:51 GMT <AFaust> And I do acknowledge the progress in terms of initial feedback. Though I have my obvious issues with the way decisions are then handled...
2017-03-03 17:03:55 GMT <resplin> My only frustration there is that I am very sympathetic to your view but don't know the solution.
2017-03-03 17:05:53 GMT <fwu> hmm.. if I set the property by code, I cna see it on the next wf screen
2017-03-03 17:05:54 GMT <resplin> I'm playing a long game. I recognize that there is going to be a lot of short term frustration. The ultimate goal is to get to GitHub, have CI testing, and have the teams empowered to pull the fixes they want to. Then the conversation is about the quality of the fix, instead of the engineering roadmap.
2017-03-03 17:06:18 GMT <fwu> but I cant get what the user wrote on the initialization screen
2017-03-03 17:06:27 GMT <resplin> JIRA makes an awful roadmap tool. If it isn't going to be done in the next quarter, it gets lost in the noise and doesn't help to have open.
2017-03-03 17:07:40 GMT <AFaust> resplin: To be clear, any frustration I have is not with you, but with Engineering. Whenever I hear/read that architects / engineers told you they think "XY would be too much effort" I could scream...
2017-03-03 17:08:33 GMT <AFaust> That team used to be efficient / capable, and it is hard for me to understand where that went. I don't want to believe that the people who left over the years pulled that much weight...
2017-03-03 17:09:06 GMT <AFaust> Or I may just not understand what it is that is draining the time / energy of the teamn
2017-03-03 17:09:34 GMT <resplin> That's the challenge of Product Management. It was easy for me to blame others until I was part of the meetings where the decisions were made, but I still disagree with many of the decisions.
2017-03-03 17:09:55 GMT <resplin> There are a few major time sucks:
2017-03-03 17:10:35 GMT <resplin> * The communication costs have grown a lot as the team has grown. Last year we had a lot of instability as we reorganized to try and address this, and I think we are only just barely seeing improvements from that.
2017-03-03 17:11:11 GMT <resplin> * We did not adequately invest in our engineering tools, like automated testing. So testing is very expensive. We have made a lot of investment there over the past year which I hope to share at BeeCon.
2017-03-03 17:11:49 GMT <resplin> * We did not adequately invest in our APIs, so it became very hard to improve the core without breaking everything. Again, we have made a lot of progress recently but aren't yet seeing the benefit.
2017-03-03 17:12:53 GMT <resplin> * We give our customer support fixes going back four or five years. And we hadn't properly budgeted to provide that benefit while still innovating on the new development. We have made some improvements here, but not enough.
2017-03-03 17:13:53 GMT <resplin> I consider this to be typical growing pains. But I understand that from the outside it looks like incompetence, indifference, or malice. I try to address that by being really transparent in what we are dealing with and how we are trying to solve things.
2017-03-03 17:16:06 GMT * AFaust is going for a sprint to the grocerie store
2017-03-03 17:17:25 GMT <resplin> AFaust: Thanks for the conversation.
2017-03-03 17:29:48 GMT <fwu> ppl, i really stuggling with this. any ideas, why I cant get the work values in the listeners of the start event? :(
2017-03-03 17:57:58 GMT <AFaust> ...and of course the board game visitors are already here so I am off for the day...
2017-03-03 18:37:53 GMT <fwu> in no way im able to get the values from my start screen arghhh
2017-03-03 18:42:48 GMT <fwu> if I set the value in a listener in the startevent I can get the value at the next task. But when the user fills in the value I cant get it.
2017-03-03 18:44:27 GMT <fwu> and I believe I dont even need to map the values in the executionListener. I believe they are set by default. But still, it doesnt work
2017-03-03 18:48:24 GMT <fwu> brb
2017-03-03 19:39:47 GMT <yreg> hello again !
2017-03-03 19:39:57 GMT <yreg> resplin, yes that was me !
2017-03-03 19:40:51 GMT <yreg> resplin, may I suggest you to use labels for better issues filtering, that way you want have to "Close" them
2017-03-03 19:41:21 GMT <yreg> I would appreciate a label "Help wanted" for example
2017-03-03 19:42:07 GMT <yreg> resplin, by the way, JIRA is an excellent tool, when used right
2017-03-03 19:43:23 GMT <yreg> Xenit invested time and money and had an expert assist on upstream/downstream meetings for a while, and now I can clearly see how that did affect (positively) the readability of dashboards and reports
2017-03-03 19:43:58 GMT <yreg> and made the use of JIRA way more effective
2017-03-03 19:44:57 GMT <yreg> it took us some time to come up with the best workflow, but it is definitely worthwhile !
2017-03-03 19:46:24 GMT <yreg> it looks like I am a bit late, and you're no longer here... any way, enjoy your weekend and feel free to bring the subject up again whenever you see fit ;-)
The other logs are at http://esplins.org/hash_alfresco