IDempiere/FullMeeting20140625 - Revision history

CarlosRuiz: drop JIRA notifications from log

2014-06-26T00:23:27Z

drop JIRA notifications from log

CarlosRuiz: full meeting

2014-06-26T00:23:03Z

full meeting

New page

⇐
[[IDempiere|Table of Contents]] |
[[IDempiere/Full Meeting Minutes|Full Meeting Minutes]] |
Full Meeting 2014-06-25


'''''CarlosRuiz''''': Good morning 
'''''tbayen''''': Daarestiet! 
'''''nmicoud''''': Bonjour ! 
'''''nmicoud''''': Carlos, could you have a look at ticket #2018 (bug with DST in calendar, hiep upload a patch that i have tested) ? 
'''''CarlosRuiz''''': good - let me check it 
'''''nmicoud''''': and right after, perhaps #2013 ? (easy patch also) 
'''''Deepak_''''': Namaste 
'''''nmicoud''''': Hi Deepak 
'''''JanThielemann''''': hello 
'''''CarlosRuiz''''': hi 
'''''Deepak_''''': Hello all 
'''''tbayen''''': Hi Deepak_ 
'''''Deepak_''''': Carlos, 1770 is waiting for your review, Hengsin has approved. 1990 you resolved but first need to import patch from 1770. 
'''''Deepak_''''': Hello Thomas 
'''''CarlosRuiz''''': Hi Deepak - reviewing 2013 at this moment .... 
'''''tbayen''''': I am interested in using more than one instance of idempiere on one single system. Does anybody have experience with that? 
'''''CarlosRuiz''''': AFAIK is just changing additional ports on tomcat xml file 
'''''tbayen''''': Yes, I managed that some times. My question is not how to do it but to share experiences. 
'''''tbayen''''': And perhaps to change somw small things like the start scripts to allow to automate it better. 
'''''tbayen''''': The only thing I am really unsure is the hazelcast cache. What happens with more instances? As we saw on the Krefeld training a number of idempiere servers in one lan can kill them all. 
'''''CarlosRuiz''''': ah yes - I think you need to set differente hazelcast port also for each instance 
'''''Not-7674''''': [IDEMPIERE] muriloht created IDEMPIERE-2020 Value Format for ZK 
'''''ocurieles_DCS''''': Hi For all my friends 
'''''Not-7674''''': [IDEMPIERE] http://idempiere.atlassian.net/browse/IDEMPIERE-2020 
'''''CarlosRuiz''''': Hi Orlando 
'''''ocurieles_DCS''''': Congratulations @CarlosRuiz for the Colombia Team :D 
'''''tbayen''''': Can we configure the hazelcast port that with a standard idempiere installation package? 
'''''CarlosRuiz''''': thanks Orlando :-) 
'''''CarlosRuiz''''': yes tbayen - there is a conf for that somewhere 
'''''CarlosRuiz''''': don't remember is it needs to be pushed as fragment or can be added directly as a property file 
'''''CarlosRuiz''''': nmicoud, how is your default timezone name on this patch? 
'''''CarlosRuiz''''': we don't have DST on Colombia - so not easy to test here :-) 
'''''nmicoud''''': hum, i have no idea, let me check 
'''''nmicoud''''': should be CET 
'''''nmicoud''''': Its name is "Europe/Paris" 
'''''nmicoud''''': as it is set automatically, i've never search for that :) 
'''''CarlosRuiz''''': ok, I see the difference that hieplq made - the old code was adding GMT+2 to the string and now must be just CET 
'''''Not-7674''''': [iDempiere] CarlosRuiz_globalqss pushed 1 commit to development [+0/-0/±1] https://bitbucket.org/idempiere/idempiere/commits/ 
'''''Not-7674''''': [iDempiere] hieplq 8316f9c - IDEMPIERE-2018:Error with DST on Calendar window 
'''''CarlosRuiz''''': tbayen, according to this info 
'''''CarlosRuiz''''': http://wiki.idempiere.org/en/NF1.0_Distributed_Cache 
'''''CarlosRuiz''''': the hazelcast.xml can be put on the idempiere home 
'''''tbayen''''': I was sure that the hazelcast communicates with the zk browser clients. That is not what is said on this page. If it is only used for load balancing isn't it the best to deactivate this bundle at all (in standard installations)? 
'''''Not-7674''''': [IDEMPIERE] carlosruiz_globalqss updated IDEMPIERE-2018 status set to "Resolved" -assignee set to "hieplq" -resolution set to "Fixed" 
'''''Not-7674''''': [IDEMPIERE] http://idempiere.atlassian.net/browse/IDEMPIERE-2018 
'''''tbayen''''': Thanks for the link. 
'''''CarlosRuiz''''': hazelcast distributes to any session 
'''''CarlosRuiz''''': be swing or another server pointing to the same database 
'''''hengsin''''': there's the group setting in hazelcast.xml 
'''''tbayen''''': But not web clients? So if you have one webserver serving some browser clients you do not need it. I would say this is the standard case, isn't it? 
'''''CarlosRuiz''''': don't remember. question for hengsin: is it worthy to have hazelcast if there is just single server? 
'''''hengsin''''': yeah, not really needed. not sure though whether there are code that's using the messaging service provided by hazelcast 
'''''hengsin''''': I means since it is always deploy, there might be code that assume it is always there ... 
'''''tbayen''''': Worth to try and deactivate it. I will tell you what happened. :-) If it does not work I will configure the hazelcast group. 
'''''tbayen''''': Thanks! 
'''''hengsin''''': what's the issue here ? 
'''''tbayen''''': I want to install more than one instance on one single machine. 
'''''hengsin''''': ic. there's no need to deactivate the hazelcast service, there just need to have different group name in hazelcast.xml 
'''''tbayen''''': The hazelcast tcp port was one of the things I identified as a possible problem. 
'''''hengsin''''': that's also configurable through hazelcast.xml 
'''''hengsin''''': <port auto-increment="true">5701</port> 
'''''hengsin''''': so by default, the next instance will auto looks for the next available port ( 5702, 5703, etc ) 
'''''Deepak_''''': tbayen, hazelcast do not communicate to zk 
'''''hengsin''''': and as long as all the instance doesn't have the same group name, they wouldn't talk to each other 
'''''tbayen''''': ok. I wondered why it worked in my tests. 
'''''Deepak_''''': But to all server instance 
'''''CarlosRuiz''''': so, you can have all instances using same port but each instance with a different group 
'''''CarlosRuiz''''': if I understood right 
'''''hengsin''''': yes, it auto looks for available port 
'''''hengsin''''': and use multicast to discover instances of the same group 
'''''Deepak_''''': hengsin, should we make hazelcast deactive by default? 
'''''tbayen''''': Thanks. I have many informations to try. :-) Thanks Carlos for the link to the wikipage. 
'''''hengsin''''': deepak, that's fine but someone will have to spend time to test it 
'''''hengsin''''': needs to announce it to the group too since it will break existing implementation 
'''''tbayen''''': To understand right: If I change the port they will nevertheless kepp in touch because they find each other through multicast. So it is needed to change the group to separate servers from each other. 
'''''tbayen''''': ? 
'''''hengsin''''': yes, that's right 
'''''hengsin''''': the port is not important, what's critical is the group name 
'''''Not-7674''''': [iDempiere] CarlosRuiz_globalqss pushed 1 commit to release-2.0 [+0/-0/±1] https://bitbucket.org/idempiere/idempiere/commits/ 
'''''Not-7674''''': [iDempiere] nmicoud 113674f - IDEMPIERE-2013 Impossible to empty EMail field on User window 
'''''hengsin''''': or if you disable multicast in your network then sure the instances can't talk to each other :) 
'''''Deepak_''''': hengsin, As I know that hazelcast is required on clustered environment and most of installation are defaultly do not need to communicate with each other 
'''''Not-7674''''': [IDEMPIERE] carlosruiz_globalqss updated IDEMPIERE-2013 status set to "Resolved" -assignee set to "Nicolas Micoud" -resolution set to "Fixed" 
'''''Not-7674''''': [IDEMPIERE] http://idempiere.atlassian.net/browse/IDEMPIERE-2013 
'''''hengsin''''': deepak, yes. but it is already there and we have make release with it 
'''''hengsin''''': now, we can't just disable it and assume it will break nothing 
'''''CarlosRuiz''''': so - the group name must be related to the db instance - could we make that a db configurable thing instead of hazelcast.xml? 
'''''Deepak_''''': So just my point is that if we have hazelcast defaulted to disable and it needs manually enable from property file 
'''''hengsin''''': carlos', not sure. most probably not. 
'''''tbayen''''': Can it be really harmful to use several servers in a lan without configuring separate groups? Will it mix up the cache and the database? I am sure many people have a testing and a producition instance. 
'''''Deepak_''''': hengsin, understood your point 
'''''hengsin''''': carlos, I guess, one possibility is we can generate hazelcast.xml that uses the db name as the group name 
'''''hengsin''''': I means that can happens as part of the setup process 
'''''Deepak_''''': hengsin, I think that is good idea 
'''''CarlosRuiz''''': the db name can be the same in different servers 
'''''CarlosRuiz''''': what I was wondering is that if you're combining swing and web - then you need a different hazelcast.xml for each swing connecting to a different instance 
'''''tbayen''''': servername_5432_dbname 
'''''CarlosRuiz''''': maybe something like this could work? 
'''''Deepak_''''': server name in group name not work 
'''''Deepak_''''': for clustered environment 
'''''CarlosRuiz''''': add line 113 on org.idempiere.hazelcast.service.Activator 
'''''CarlosRuiz''''': config.getGroupConfig().setName(dbserver + dbport + dbname); 
'''''CarlosRuiz''''': ah - or as hengsin suggested - for webserver we can generate the hazelcast.xml with that group name - and for swing client it can set the name programatically 
'''''tbayen''''': For servers it can be the same as for swing. It seems the same problem. 
'''''tbayen''''': You can check if there is a hazelcast.xml in the filesystem. If not we set the group name programatically. 
'''''Not-7674''''': [IDEMPIERE] carlosruiz_globalqss updated IDEMPIERE-1770 
'''''Not-7674''''': [IDEMPIERE] [~deepak], peer review: 1 - About migration/i2.0z/oracle/201404170743_IDEMPIERE-1770.sql migration/i2.0z/oracle/201404170743_IDEMPIERE-1770.sql db/ddlutils/oracle/views/M_InoutLineMA_Returned.sql db/ddlutils/postgresql/views/M_InoutLineMA_Returned.sql I think is not good idea to have a "SELECT sma.*" in a view as each installation can have different columns and order of columns for a table. better to explicitly set the columns required on 
'''''Not-7674''''': [IDEMPIERE] http://idempiere.atlassian.net/browse/IDEMPIERE-1770 
'''''CarlosRuiz''''': muriloht, AYT? 
'''''CarlosRuiz''''': muriloht, sent you the centralized ID password to your emails 
'''''Not-7674''''': [IDEMPIERE] carlosruiz_globalqss updated IDEMPIERE-2020 
'''''Not-7674''''': [IDEMPIERE] [~muriloht], just sent centralized ID password to your email. 
'''''Not-7674''''': [IDEMPIERE] http://idempiere.atlassian.net/browse/IDEMPIERE-2020 
'''''muriloht''''': hi Carlos, i'm here now... 
'''''CarlosRuiz''''': I found two emails from you :-) sent password to both 
'''''muriloht''''': I received the email with my login for the centralized id, thank you! 
'''''CarlosRuiz''''': you're welcome 
'''''muriloht''''': tks! iDempiere is an amazing project! i hope to contribute and learn as much as possible from now on 
'''''CarlosRuiz''''': need to go out for 2 hours - talk to you later 
'''''tbayen''''': bye 
'''''tbayen''''': Hi muriloht, welcome to the club! Where are you from? 
'''''muriloht''''': hi tbayen, i'm from Brazil! 
'''''muriloht''''': (São Paulo) 
'''''CarlosRuiz''''': change of plans - appointment cancelled 
'''''CarlosRuiz''''': will take a look to some tickets from hieplq 
'''''hieplq_''''': hi CarlosRuiz. i need your suport in some issue. please comment your idea to help me. 
'''''hieplq_''''': IDEMPIERE-1991 
'''''hieplq_''''': IDEMPIERE-1989 
'''''hieplq_''''': IDEMPIERE-1883 
'''''hieplq_''''': IDEMPIERE-2019 
'''''hieplq_''''': get your last message, many thanks. i'm here to support you in my work :) 
'''''Stanley_''''': Hi 
'''''Stanley_''''': All 
'''''Stanley_''''': hi red1 
'''''Stanley_''''': Hi @red1 
'''''Not-7674''''': [IDEMPIERE] carlosruiz_globalqss updated IDEMPIERE-2019 
'''''Not-7674''''': [IDEMPIERE] [~hieplq], I think that's ok, by default the info window just show customers (because of the sotrx flag) but the field allows any BP. The effect would be similar if you set a default on any info window but that default doesn't restrict the referenced table. Regards, Carlos Ruiz 
'''''Not-7674''''': [IDEMPIERE] http://idempiere.atlassian.net/browse/IDEMPIERE-2019 
'''''Not-7674''''': [IDEMPIERE] carlosruiz_globalqss updated IDEMPIERE-1883 
'''''Not-7674''''': [IDEMPIERE] Seems like a bug - the tree must show the records associated to the header. 
'''''Not-7674''''': [IDEMPIERE] http://idempiere.atlassian.net/browse/IDEMPIERE-1883 
'''''Not-7674''''': [IDEMPIERE] carlosruiz_globalqss updated IDEMPIERE-1883 
'''''Not-7674''''': [IDEMPIERE] Seems like a bug - the tree must show the records associated to the header. Or maybe the cause is that you must not have two elements pointing to the same tree? 
'''''Not-7674''''': [IDEMPIERE] http://idempiere.atlassian.net/browse/IDEMPIERE-1883 
'''''Not-7674''''': [IDEMPIERE] carlosruiz_globalqss updated IDEMPIERE-1989 
'''''Not-7674''''': [IDEMPIERE] [~hieplq], there are just two special cases where a process have duplicated columnname in process paramters: - Copy Role - Mutiple Components Change (inactive from libero mfg) I think that's poor design and would be better to forbid that and refactor the copy role process, and maybe the libero mfg process on red1's plugin. Regards, Carlos Ruiz 
'''''Not-7674''''': [IDEMPIERE] http://idempiere.atlassian.net/browse/IDEMPIERE-1989 
'''''Not-7674''''': [IDEMPIERE] carlosruiz_globalqss updated IDEMPIERE-1991 
'''''Not-7674''''': [IDEMPIERE] as a workaround maybe you could have a virtual column on the tab using the logic expression? which is the use case of this one? which logic do you refer (display, mandatory, readonly)? 
'''''Not-7674''''': [IDEMPIERE] http://idempiere.atlassian.net/browse/IDEMPIERE-1991 
'''''nmicoud''''': CarlosRuiz, hiep uplaod a patch for ticket #1899 (broken image in email signature). Could you have a look at it and integrate it ? I've tested it and it seems ok. 
'''''CarlosRuiz''''': ok 
'''''nmicoud''''': thanks 
'''''CarlosRuiz''''': nmicoud, hieplq_ in my tests the method replaceBASE64Img is not finding the string to replace back 
'''''nmicoud''''': mean that the mail is sent not correctly ? 
'''''CarlosRuiz''''': is sent correctly - but the images are duplicated - so the size is bigger 
'''''nmicoud''''': i haven't noticed that. You use same mail text ? 
'''''CarlosRuiz''''': I noticed looking the message source in thunderbird 
'''''nmicoud''''': ah ok, i haven't do that 
'''''nmicoud''''': so patch is not final 
'''''CarlosRuiz''''': trying to find the reason 
'''''nmicoud''''': ok, thanks 
'''''CarlosRuiz''''': it worked if I didn't change the message 
'''''CarlosRuiz''''': but if I change the message then the problem happens 
'''''nmicoud''''': "change message" = modifying the content of signature or other text ? 
'''''CarlosRuiz''''': modify the content of the mail before pushing ok 
'''''CarlosRuiz''''': the CKeditor drops the srcSave="" string which hieplq_ used to preserve the original name of the image 
'''''CarlosRuiz''''': ah! easy workaround we can use alt-text 
'''''CarlosRuiz''''': testing ... 
'''''nmicoud''''': good idea ! 
'''''CarlosRuiz''''': yes - it worked - I'm going to add a validation that the image is an attachment before replacing 
'''''nmicoud''''': great ! 
'''''Not-7674''''': [iDempiere] CarlosRuiz_globalqss pushed 2 commits to development [+0/-0/±1] https://bitbucket.org/idempiere/idempiere/commits/ 
'''''Not-7674''''': [iDempiere] hieplq bd22c80 - IDEMPIERE-1899:EMail : Implement default mail text and allow images within message 
'''''Not-7674''''': [iDempiere] globalqss 3a1a6de - IDEMPIERE-1899 EMail : Implement default mail text and allow images within message 
'''''nmicoud''''': thanks a lot Carlos, will put it in our test instance and give feedback 
'''''Not-7674''''': [IDEMPIERE] carlosruiz_globalqss updated IDEMPIERE-1899 status set to "Resolved" -resolution set to "Fixed" 
'''''Not-7674''''': [IDEMPIERE] Thanks [~hieplq] [~nmicoud], effectively the problem you described is a thunderbird bug: https://bugzilla.mozilla.org/show_bug.cgi?id=548507 I tested with thunderbird and didn't forward not-inline attachments, tested from within gmail and it worked fine. Regards, Carlos Ruiz 
'''''Not-7674''''': [IDEMPIERE] http://idempiere.atlassian.net/browse/IDEMPIERE-1899

@@ Line 28: / Line 28: @@
 '''''tbayen''''': The only thing I am really unsure is the hazelcast cache. What happens with more instances? As we saw on the Krefeld training a number of idempiere servers in one lan can kill them all.<br>
 '''''CarlosRuiz''''': ah yes - I think you need to set differente hazelcast port also for each instance<br>
 '''''ocurieles_DCS''''': Hi For all my friends<br>
 '''''CarlosRuiz''''': Hi Orlando<br>
 '''''ocurieles_DCS''''': Congratulations @CarlosRuiz for the Colombia Team :D<br>
@@ Line 44: / Line 42: @@
 '''''nmicoud''''': as it is set automatically, i've never search for that :)<br>
 '''''CarlosRuiz''''': ok, I see the difference that hieplq made - the old code was adding GMT+2 to the string and now must be just CET<br>
 '''''CarlosRuiz''''': tbayen, according to this info<br>
 '''''CarlosRuiz''''': http://wiki.idempiere.org/en/NF1.0_Distributed_Cache<br>
 '''''CarlosRuiz''''': the hazelcast.xml can be put on the idempiere home<br>
 '''''tbayen''''': I was sure that the hazelcast communicates with the zk browser clients. That is not what is said on this page. If it is only used for load balancing isn't it the best to deactivate this bundle at all (in standard installations)?<br>
 '''''tbayen''''': Thanks for the link.<br>
 '''''CarlosRuiz''''': hazelcast distributes to any session<br>
@@ Line 85: / Line 79: @@
 '''''hengsin''''': yes, that's right<br>
 '''''hengsin''''': the port is not important, what's critical is the group name<br>
 '''''hengsin''''': or if you disable multicast in your network then sure the instances can't talk to each other :)<br>
 '''''Deepak_''''': hengsin, As I know that hazelcast is required on clustered environment and most of installation are defaultly do not need to communicate with each other<br>
 '''''hengsin''''': deepak, yes. but it is already there and we have make release with it<br>
 '''''hengsin''''': now, we can't just disable it and assume it will break nothing<br>
@@ Line 112: / Line 102: @@
 '''''tbayen''''': For servers it can be the same as for swing. It seems the same problem.<br>
 '''''tbayen''''': You can check if there is a hazelcast.xml in the filesystem. If not we set the group name programatically.<br>
 '''''CarlosRuiz''''': muriloht, AYT?<br>
 '''''CarlosRuiz''''': muriloht, sent you the centralized ID password to your emails<br>
 '''''muriloht''''': hi Carlos, i'm here now...<br>
 '''''CarlosRuiz''''': I found two emails from you  :-)   sent password to both<br>
@@ Line 142: / Line 126: @@
 '''''Stanley_''''': hi red1<br>
 '''''Stanley_''''': Hi @red1<br>
 '''''nmicoud''''': CarlosRuiz, hiep uplaod a patch for ticket #1899 (broken image in email signature). Could you have a look at it and integrate it ? I've tested it and it seems ok.<br>
 '''''CarlosRuiz''''': ok<br>
@@ Line 179: / Line 148: @@
 '''''CarlosRuiz''''': yes - it worked - I'm going to add a validation that the image is an attachment before replacing<br>
 '''''nmicoud''''': great !<br>
 '''''nmicoud''''': thanks a lot Carlos, will put it in our test instance and give feedback<br>