Daily Log for #alfresco IRC Channel

Alfresco discussion and collaboration. Stick around a few hours after asking a question.

Official support for Enterprise subscribers: support.alfresco.com.

Joining the Channel:

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.

Getting Help

More help is available in this list of resources.

Daily Log for #alfresco

2018-02-19 07:20:57 GMT <Younes> Morning everyone !

2018-02-19 07:58:49 GMT <twen> hello

2018-02-19 07:59:47 GMT <ayushi> hi all,

2018-02-19 07:59:49 GMT <Younes> hello twen !

2018-02-19 08:00:01 GMT <Younes> how it goes pal ?

2018-02-19 08:00:14 GMT <ayushi> can anyone tell me if there is any api to get the uuid from noderef in alfresco

2018-02-19 08:06:17 GMT <twen> hey Younes :) I'm fine thanks. Back in Switzerland ^^

2018-02-19 08:08:08 GMT <Younes> ayushi, noderef.getId()

2018-02-19 08:08:14 GMT *** Younes is now known as yreg

2018-02-19 09:41:19 GMT <Tichodroma> Do you use beans derived from basePatch to add DB tables? How do you handle the problem that the targetSchema property of such a bean must be larger than all official DB patches? Just set it to 99999?

2018-02-19 09:46:05 GMT <angelborroy> Jeff is using 15,000

2018-02-19 09:46:06 GMT <angelborroy> https://github.com/jpotts/share-site-creators/blob/master/share-site-creators-repo/src/main/resources/alfresco/module/share-site-creators-repo/context/bootstrap-context.xml#L26

2018-02-19 09:46:07 GMT <alfbot> Title: share-site-creators/bootstrap-context.xml at master · jpotts/share-site-creators · GitHub (at github.com)

2018-02-19 09:46:22 GMT <angelborroy> Tichodroma It looks like Alfresco will not get that number in years

2018-02-19 09:47:08 GMT <Tichodroma> :)

2018-02-19 09:47:18 GMT <Tichodroma> Then I'll use 25,000

2018-02-19 09:47:56 GMT <angelborroy> anyway that “adding DB tables to schema” makes me feeling nervous…

2018-02-19 09:48:11 GMT <angelborroy> … but I know you knows, so go ahead!

2018-02-19 09:48:12 GMT <Loftux> Think there needs to be a bidding war, who will pick the highest number ;)

2018-02-19 09:51:19 GMT * DarkStar1 bids you all a good morning

2018-02-19 10:07:49 GMT <Tichodroma> angelborroy: Loftux: I'd prever if there was a more "official" way to manage custom tables.

2018-02-19 10:09:24 GMT <angelborroy> I’m creating custom tables outside Alfresco database

2018-02-19 10:09:34 GMT <Tichodroma> how?

2018-02-19 10:09:50 GMT <angelborroy> ?

2018-02-19 10:09:53 GMT <Tichodroma> I am looking for a way to do this from a module

2018-02-19 10:10:08 GMT <angelborroy> I’m using a JDBC connection from inside the module

2018-02-19 10:10:14 GMT <angelborroy> what else?

2018-02-19 10:10:20 GMT <angelborroy> Trasnationality?

2018-02-19 10:10:44 GMT <Tichodroma> so you don't use a bean derived from public abstract class AbstractPatch implements Patch, ApplicationEventPublisherAware

2018-02-19 10:10:51 GMT <Tichodroma> argh

2018-02-19 10:11:03 GMT <Tichodroma> so you don't use a bean derived from AbstractPatch?

2018-02-19 10:11:13 GMT <angelborroy> for custom tables, no

2018-02-19 10:11:32 GMT <Tichodroma> could be an idea

2018-02-19 10:12:26 GMT <angelborroy> you need to listen to Alfresco Transaction to perform a coordinated commit/rollback

2018-02-19 10:13:17 GMT <angelborroy> I’m using default TransactionListener service for that

2018-02-19 10:29:45 GMT *** angelborroy_ is now known as angelborroy

2018-02-19 10:31:21 GMT <AFaust> Tichodroma, angelborroy: You could always just use the placeholder ${version.schema} for the targetSchema, and you should always be safe...

2018-02-19 10:32:55 GMT <AFaust> A customer of mine just added a 0 to the previous default hard-coded 10000 on the same assumption that Alfresco won't reach that number...and they did not know that any property from version.properties could be referenced in Spring

2018-02-19 10:33:30 GMT <Tichodroma> AFaust: fixesToSchema is ${version.schema}, would you use this for targetSchema, too?

2018-02-19 10:33:58 GMT <AFaust> angelborroy: Why not use the same MyBatis/iBatis template structure and have the Alfresco transaction handling rollback/commit your stuff?...

2018-02-19 10:34:36 GMT <angelborroy> This can be done having 2 differents JDBC sources?

2018-02-19 10:34:43 GMT <angelborroy> I didn’t knew thata

2018-02-19 10:35:03 GMT <AFaust> Tichodroma: For custom SQL patches, yes. Though I very rarely write patches - I prefer ModuleComponent for extensions..

2018-02-19 10:35:22 GMT <Tichodroma> AFaust: legacy ...

2018-02-19 10:36:03 GMT <AFaust> angelborroy: Well, for different JDBC sources it is a bit more complex, but you can also add the same kind of AOP handling that Alfresco uses for transaction demarcation...

2018-02-19 10:36:22 GMT <angelborroy> TransationListener is not a right approach?

2018-02-19 10:36:25 GMT <angelborroy> It works for me

2018-02-19 10:37:03 GMT <AFaust> Sure it works.... depends on how you handle the listener registration. If it is done in a way that there is no chance of it being "forgotten" by a careless dev...

2018-02-19 10:37:15 GMT <AFaust> That's why I prefer AOP for stuff like this...

2018-02-19 10:37:36 GMT <AFaust> And the MyBatis/iBatis stuff to make it (mostly) DB / dialect independant

2018-02-19 11:02:17 GMT <Sumit> any standard configs to enable sign in with email/pass pairs?

2018-02-19 11:03:19 GMT <Tichodroma> Sumit: no

2018-02-19 11:05:08 GMT <Sumit> ok thanks. hope this good to have because email is also a unique attribute.

2018-02-19 11:05:48 GMT <Tichodroma> we've implemented this as a module

2018-02-19 11:06:30 GMT <Tichodroma> so if you are interested, feel free to contact us: http://www.ecm4u.de/en/contact.php

2018-02-19 11:06:31 GMT <alfbot> Title: Contakt (at www.ecm4u.de)

2018-02-19 11:07:40 GMT <AFaust> Sumit: If your users are in LDAP/AD and you only want to provide email + PW login for those (without simple user name support), this could be achieved by mapping the email attribute as the user name with standard Alfresco

2018-02-19 11:08:23 GMT <AFaust> Solutions such as those from ecm4u primarily provide the additional flexibility of fallbacks / alternative login methods being supported at the same type

2018-02-19 11:08:25 GMT <Sumit> Yes Afaust

2018-02-19 11:09:29 GMT <Sumit> But I was trying for dms user.

2018-02-19 11:09:35 GMT <Tichodroma> but be aware that this LDAP attribute mapping can't be changed or you will delete and recreate all users

2018-02-19 11:09:39 GMT <Sumit> Yes with LDAP it's possible

2018-02-19 11:09:42 GMT <AFaust> ... in some cases I guess you could also support simple login + email login with standard Alfresco using the user-name resolution feature of the LDAP subsystem...

2018-02-19 11:10:03 GMT <AFaust> +1 to Tichodroma if you have already synchronized the users...

2018-02-19 11:10:08 GMT <Tichodroma> there are some gotchas to take care of in Share

2018-02-19 11:12:12 GMT <AFaust> And of course these gotchas depend on the Alfresco version. When I worked on my alfresco-mt-support LDAP module for 5.1+ (allowing some alternative resolutions) I did not encounter gotchas...

2018-02-19 14:22:06 GMT <yreg> AFaust, We have a usecase where all content needs to be moved from one store to one other (different store types); And a colleague of mine thinks the best way to get that done, is to iterate over all content_url/content_data items and "touch" them after configuring stores to default to the target content store and have the source (old) chained

2018-02-19 14:23:30 GMT <yreg> I was wondering if that would actually affect solr and eventually trigger a full reindex and if there were better options

2018-02-19 14:23:44 GMT <yreg> have you run into such requirements yourself ?

2018-02-19 14:23:50 GMT <AFaust> Well, there is no simple "touch" for Content URL / Data data structures that would move content. You'd have to write again to the actual node property...

2018-02-19 14:25:35 GMT <AFaust> I have run across simple move scenarios between the same type of store, e.g. from one file content store to another

2018-02-19 14:27:08 GMT <AFaust> and such a thing is simple

2018-02-19 14:28:07 GMT <yreg> in the case it is a migration from file content store to caringo storage content store

2018-02-19 14:28:28 GMT <AFaust> What I would do is a two stage approach...

2018-02-19 14:29:30 GMT <AFaust> 1) configure a store setup with old store as primary and new store as secondary store, where read-access would replicate content from primary to secondary - then iterate over all content_url items and simply retrieve a reader from the content service and access its content (while not actually doing anything with the bytes)

2018-02-19 14:30:25 GMT <AFaust> 2) configure the actual target store(s)

2018-02-19 14:30:39 GMT <AFaust> The CachingContentStore would be the best option for 1)

2018-02-19 14:31:04 GMT <AFaust> Since the ReplicatingContentStore has been removed in 5.2...

2018-02-19 14:36:26 GMT <DarkStar1> Qq, anyone. Is there an AOS session/connection timeout property? We have a client that's telling us that MS word prompts for auth and will offer to save an open document as a new version after some minutes of inactivity.

2018-02-19 14:37:11 GMT <DarkStar1> Sorry my cold isn't making me think. By new version I mean save the document under a different name

2018-02-19 14:39:52 GMT <yreg> AFaust, I haven't had much contact with content stores, so I am not really following up on your advice, I will however share it as-is with my colleague, thanks !

2018-02-19 14:41:57 GMT <yreg> DarkStar1, AFAIK it is supposed to be using basic authentication so I find it weird to have to put in password again..

2018-02-19 14:44:56 GMT <DarkStar1> yreg: Apologies. He says after a few minutes of inactivity he gets a save as dialog for the document and has to save it under a different name. Not an auth.

2018-02-19 14:58:35 GMT <DarkStar1> Going to crash for the day. I'll look into this issue tomorrow when I should hopefully be better.

2018-02-19 16:33:35 GMT <mbui> I have this issue with solr indexes continously getting more nodes that are marked as "Alfresco Error Nodes in Index" which I get by executing the “solr4/admin/cores?action=SUMMARY&wt=xml”. I have 5% of all nodes marked as errors (95k out of 1.8m nodes). How do I solve this and prevent it from happening in the future? The most recent reindexing was 2 months ago. Is it possible to fix it without reindexing?

2018-02-19 16:35:37 GMT <mbui> I seem to be getting alot of these in the logs:

2018-02-19 16:35:44 GMT <mbui> ERROR [node.ibatis.NodeDAOImpl$MySQL] [http-bio-8043-exec-1] Cyclic parent-child relationship detected: … ... Caused by: org.alfresco.service.cmr.repository.CyclicChildRelationshipException: Node has been pasted into its own tree. .. .. org.alfresco.repo.solr.SOLRTrackingComponentImpl.getNodesMetadata(SOLRTrackingComponentImpl.java:872)

2018-02-19 16:38:23 GMT <mbui> I'm getting a lot of these as well: INFO [repo.sync.SyncTrackerComponent] [SyncTrackerPush3] PUSH: pushed syncSet '33e6e44b-599f-4f48-8fb0-7da16b721d90' nodes (0 succeeded, 37 failed) [in 79 ms]

2018-02-19 17:25:27 GMT *** alfbot` is now known as alfbot

2018-02-19 18:27:33 GMT <AFaust> mbui: Are you using Alfresco Enterprise Edition + Cloud Sync? Looks like somehow the sync is creating a cycle where the child node of another (or any other descendant) also appears as a parent node (or an ancestor)

2018-02-19 18:28:45 GMT <AFaust> Well - could have been created by something else as well (not just cloud sync)... Though there is generally pretty decent protection against that, so it must be either some low-level process or someone is modifying the DB without using Alfresco services...

2018-02-19 18:32:02 GMT <mbui> AFaust: Yes, we have Enterprise and Cloud sync enabled.

2018-02-19 18:32:18 GMT <mbui> I also seem to be getting a few of these if they are related to the issue:

2018-02-19 18:32:18 GMT <mbui> ERROR [util.transaction.TransactionSupportUtil] [defaultAsyncAction4] After completion (committed) TransactionalCache exception org.alfresco.error.AlfrescoRuntimeException: 011939112 Failed to extract audit data:

2018-02-19 18:32:51 GMT <AFaust> So, for the SyncTrackerPush job open a ticket with Support. That Cloud sync stuff has so many potential error causes that various customers of mine have given up using it...

2018-02-19 18:33:59 GMT <AFaust> That last error is so generic, it can affect anything...

2018-02-19 18:34:02 GMT <mbui> Alright, thanks i'll do that. Any tips on the information the support may require which will help the investigation? Except catalina.out logs?

2018-02-19 18:34:31 GMT <AFaust> Same thing as always, full logs (alfresco.log / catalina.log), JMX dump etc. - the whole shebang

2018-02-19 18:34:49 GMT <mbui> Right, thank you master Faust.

2018-02-19 18:34:58 GMT <AFaust> And then have fun doing trial&error analysis with their feedback...

2018-02-19 18:35:26 GMT <AFaust> Cloud sync was one of the most annoying bits to try to analyse / solve with Support.

2018-02-19 18:36:18 GMT <mbui> So you believe that the 5% of my nodes having corrupted indexes is tied with cloud sync?

2018-02-19 18:36:56 GMT <mbui> That % is increasing every second by the way.

2018-02-19 18:37:11 GMT <mbui> Hmm. minute rather.

2018-02-19 18:37:49 GMT <AFaust> I don't know. You just posted the errrors so close together that I assumed you already correlated them...

2018-02-19 18:38:30 GMT <AFaust> Without knowing the details of your system and customisations I could only do wild guesses

2018-02-19 18:39:09 GMT <AFaust> I mean, what are the EXCEPTIONMESSAGE entries you have in SOLR? (I presume you already checked this...)

2018-02-19 18:39:40 GMT <AFaust> Can be transformation issues, timeouts...etc. not just the one CyclicChildRelationshipException you already listed...

2018-02-19 18:43:16 GMT <mbui> Yeah, it's a nightmare to pin point. Thanks for the help though.

End of Daily Log

The other logs are at http://esplins.org/hash_alfresco