This is still happening when I shut down and restart tomcat fresh, and then run the initialization from the script instead of the browser.

I’m sure I’ve done something simple wrong, but I can’t figure out what.

From: Ray Lee [mailto:rhlee@berkeley.edu]
Sent: Friday, April 29, 2016 4:54 PM
To: RAY, Eric Eric.Ray@victoriacollege.edu
Cc: Aron Roberts aron@socrates.berkeley.edu; talk@lists.collectionspace.org
Subject: Re: [Talk] ant build.xml error on install

I've seen the ConcurrentUpdateException happen when two initializations get kicked off simultaneously, which is usually because your browser decides to prefetch the authorities/initialise page as you're typing in the URL. Then you hit enter, and it fetches again while the first fetch is still going. It's CollectionSpace's fault for violating the convention that GET requests are for read (not write) operations. You can diagnose this after the fact by looking in the tomcat access log to see if two GETs were received when you only expected one.

Sometimes two initializations will interleave in such a way that they both succeed, and you end up with duplicated entries.

Usually I end up tailing the cspace-app log while typing in the URL, and if stuff starts happening, I don't hit enter. It's awkward, but at least you don't have to initialize authorities very often.

Ray

On Fri, Apr 29, 2016 at 2:20 PM, RAY, Eric <Eric.Ray@victoriacollege.edumailto:Eric.Ray@victoriacollege.edu> wrote:
More issues! I cloned the core tenant using the ‘easy way’ described in the documentation. That appears to have worked, and I can log in.

When I try to initialize either with the URL or the script, there’s a failure.

Here’s what I have from the browser.

testing Authority  vocab
need to create Instance vocab-xxx
that was weird but probably not a problem Connection exceptionorg.nuxeo.ecm.core.api.ConcurrentUpdateException: Failed to save session: org.nuxeo.ecm.core.api.ConcurrentUpdateDocumentException: org.nuxeo.ecm.core.storage.ConcurrentUpdateStorageException: java.sql.BatchUpdateException: Batch entry 0 INSERT INTO "vocabularies_common" ("id", "vocabtype", "shortidentifier", "description", "source", "displayname", "refname") VALUES ('ffacf4c5-0e0a-4c8a-8f1a-85bb2c3e7c27', NULL, 'xxx', NULL, NULL, 'Temporary Vocab whilst none exist', 'urn:cspace:museumofthecoastalbend.org:vocabularies:name(xxx)''Temporary Vocab whilst none exist''') was aborted.  Call getNextException to see the cause.

I tried cloning a new tenant and trying again – same issue.

From: aronroberts@gmail.commailto:aronroberts@gmail.com [mailto:aronroberts@gmail.commailto:aronroberts@gmail.com] On Behalf Of Aron Roberts
Sent: Friday, April 29, 2016 3:01 PM

To: RAY, Eric <Eric.Ray@victoriacollege.edumailto:Eric.Ray@victoriacollege.edu>
Cc: Richard Millet <richard.millet@lyrasis.orgmailto:richard.millet@lyrasis.org>; talk@lists.collectionspace.orgmailto:talk@lists.collectionspace.org
Subject: Re: [Talk] ant build.xml error on install

On Fri, Apr 29, 2016 at 12:46 PM, RAY, Eric <Eric.Ray@victoriacollege.edumailto:Eric.Ray@victoriacollege.edu> wrote:
While the csadmin user was set correctly in pg_hba.conf, the localhost entry was set to ident instead of trust.  I don’t know why – must have come default. I changed it to trust, and the build worked. I can now log in with the default credentials.

Onward and upward…

Nice - thanks for reporting on your findings here!

The automated installer didn’t work when I tried it – I’m on CentOS. I can’t remember what the issue was, but I just went ahead with the regular installation process.

Makes sense. The CollectionSpace v4.3 installer has been tested with Ubuntu 14.04 LTS, but not so far with other Linux distributions or releases.
We've done a successful install using the v4.2 installer on CentOS 7, while there's a known issue installing on CentOS 6.x:

https://issues.collectionspace.org/browse/CSPACE-6518
And despite that success, it looks there's still some work remaining to do for the OpenJDK install in CentOS 7:

https://issues.collectionspace.org/browse/CSPACE-6721

I'm hoping at some point that installer can be regularly updated to work with (at least) the then-current, long-term support releases of both Ubuntu and CentOS.
Thanks,
Aron

From: aronroberts@gmail.commailto:aronroberts@gmail.com [mailto:aronroberts@gmail.commailto:aronroberts@gmail.com] On Behalf Of Aron Roberts
Sent: Friday, April 29, 2016 1:41 PM

To: RAY, Eric <Eric.Ray@victoriacollege.edumailto:Eric.Ray@victoriacollege.edu>
Cc: Richard Millet <richard.millet@lyrasis.orgmailto:richard.millet@lyrasis.org>; talk@lists.collectionspace.orgmailto:talk@lists.collectionspace.org
Subject: Re: [Talk] ant build.xml error on install

Hi Eric,

This could potentially mean that, in PostgreSQL's host-based authentication file, pg_hba.conf, the authentication method for the 'csadmin' user is set to 'ident', which would require that database access requests come from an operating system user with that same name as that PostgreSQL user ('csadmin'). To resolve this, the relevant entry in that configuration file should likely be changed from 'ident' to 'md5', to allow (hashed) password-based authentication.

For instructions, including an example pg_hba.conf file, see the "Changes to pg_hba.conf" section in this document:
https://wiki.collectionspace.org/display/DOC/PostgreSQL+Installation+under+Linux#PostgreSQLInstallationunderLinux-Changestopg_hba.conf

And as background, there's a description of PostgreSQL's various authentication methods here:
http://www.postgresql.org/docs/current/static/auth-pg-hba-conf.html

BTW, have you had a chance yet to try out the automated CSpace installer? We'd love to have your feedback on it:
https://wiki.collectionspace.org/display/DOC/Automated+installer+for+CollectionSpace+4.3
The goal of that installer is to allow entering a couple of commands, then walk away for about 30 minutes, and come back to a (nearly) complete installation, with just a few finishing steps to perform. There are a lot of steps in a CSpace installation, and we'd like for this to be far less painful.

Thanks,
Aron

On Fri, Apr 29, 2016 at 11:11 AM, RAY, Eric <Eric.Ray@victoriacollege.edumailto:Eric.Ray@victoriacollege.edu> wrote:
As it happens, I do have another issue that’s just come up.

I had gotten the login page running, but realized that logins (even the default core login) were failing, and perhaps that had to do with outdated postgres. So I upgraded postgres to 9.5, erased the tomcat and collectionspace-source, and started over. All went fine, except that in the ant undeploy deploy create_db -Drecreate_db=true import command, I’m getting the following error:

BUILD FAILED
/home/mcb/collectionspace-source/services/build.xml:220: The following error occurred while executing this line:
/home/mcb/collectionspace-source/services/build.xml:256: org.postgresql.util.PSQLException: FATAL: Ident authentication failed for user "csadmin"

I believe that csadmin is set up in postgres correctly, and I believe that all my environment variables are set. I can echo the environment variables for passwords, and they match what I set the csadmin password to be.

Any thoughts?

From: aronroberts@gmail.commailto:aronroberts@gmail.com [mailto:aronroberts@gmail.commailto:aronroberts@gmail.com] On Behalf Of Aron Roberts
Sent: Friday, April 29, 2016 1:03 PM
To: RAY, Eric <Eric.Ray@victoriacollege.edumailto:Eric.Ray@victoriacollege.edu>
Cc: Richard Millet <richard.millet@lyrasis.orgmailto:richard.millet@lyrasis.org>; talk@lists.collectionspace.orgmailto:talk@lists.collectionspace.org
Subject: Re: [Talk] ant build.xml error on install

Thanks much, Eric, for this update. Glad that you were able to get past this stuck point! Appreciate your trying out CollectionSpace, and please don't hesitate to ask any other questions that come up during the process.
In addition to what Richard mentioned, one other possibility that came to mind was file access permissions, for various files/directories in the Tomcat server folder. I've run into those issues when expanding that folder tarball and/or when running some install commands, as root (using 'sudo'), and then running some later commands as an unprivileged user. That might not have been applicable in this case, but am mentioning this in case it helps anyone going forward ...
Aron Roberts
UC Berkeley

On Fri, Apr 29, 2016 at 6:44 AM, RAY, Eric <Eric.Ray@victoriacollege.edumailto:Eric.Ray@victoriacollege.edu> wrote:
This error, for future reference, must have come from an error of mine in previous steps. I started over and it wasn't present the second time around.

Thanks!

On Apr 28, 2016, at 6:35 PM, Richard Millet <richard.millet@lyrasis.orgmailto:richard.millet@lyrasis.org> wrote:

Ray,

Just a guess but this error message seems to indicate that the file (or its parent directory) is read-only or perhaps in-use?  Maybe the tomcat server is still running?

-Richard

From: Talk <talk-bounces@lists.collectionspace.orgmailto:talk-bounces@lists.collectionspace.org> on behalf of RAY, Eric <Eric.Ray@victoriacollege.edumailto:Eric.Ray@victoriacollege.edu>
Sent: Thursday, April 28, 2016 4:19 PM
To: talk@lists.collectionspace.orgmailto:talk@lists.collectionspace.org
Subject: [Talk] ant build.xml error on install

Hello,

I’m working my way through a CollectionSpace install.

When I run the ant undeploy deploy create_db -Drecreate_db=true import command, I get the following errors. At first ant was out of date, but I updated it to 1.9.7, and this is verified by ant –version.

BUILD FAILED

/home/mcb/collectionspace-source/services/build.xml:313: The following error occurred while executing this line:

/home/mcb/collectionspace-source/services/services/build.xml:235: The following error occurred while executing this line:

/home/mcb/collectionspace-source/services/services/common/build.xml:183: The following error occurred while executing this line:

/home/mcb/collectionspace-source/services/services/common/build.xml:199: Unable to delete file /usr/local/share/apache-tomcat-7.0.57/lib/spring-security-acl-3.0.5.RELEASE.jar

Any thoughts?

Eric Ray

Curator of Collections and Exhibits

Museum of the Coastal Bend at Victoria College

2200 E. Red River, Victoria, TX 77901

361-572-6468tel:361-572-6468 | eric.ray@victoriacollege.edumailto:eric.ray@victoriacollege.edu

Talk mailing list
Talk@lists.collectionspace.orgmailto:Talk@lists.collectionspace.org
http://lists.collectionspace.org/mailman/listinfo/talk_lists.collectionspace.org

Talk mailing list
Talk@lists.collectionspace.orgmailto:Talk@lists.collectionspace.org
http://lists.collectionspace.org/mailman/listinfo/talk_lists.collectionspace.org

Hi Eric,

One other thought: there's a known issue - just recently surfaced - with
the "easy way":

https://issues.collectionspace.org/browse/CSPACE-6944

If you like, you could try 'cloning' that file from the 'core' tenant,
and renaming it and changing its contents via Step 6-on, per these
instructions:

https://wiki.collectionspace.org/display/DOC/Creating+your+new+tenant#Creatingyournewtenant-CreateandmodifytheServiceslayerfilesandfoldersforyournewtenant

(I haven't experimented with this possible workaround, but it's worth a
try, before going the longer, manual route ...)

Aron

On Fri, Apr 29, 2016 at 3:02 PM, RAY, Eric Eric.Ray@victoriacollege.edu
wrote:

Eric,

I'm really sorry you're still having trouble with the install.  Here are a couple other things to try:

1. Shutdown tomcat and manually delete any and all CollectionSpace related databases that were created in your postgres server.  If you haven't changed the CSPACE_INSTANCE environment variable, all the databases created by CollectionSpace should end with "_default".  You can also delete all the associated login roles that also end with "_default".  I use the pgAdmin tool to do these type of tasks, but you can also do this using the psql command line tool.

2. From the "services" layer's root directory, run "mvn clean install -DskipTests"

3. From the "application" layer's root source directory, run "mvn clean install -DskipTests"

4. Back to the "services" layer's root directory, run "ant undeploy deploy create_db -Drecreate_db=true import"

5. Restart tomcat

6. Make sure the URL of the initialization script has the correct tenant ID and that the script has the correct tenant credentials.  For example, the core/default tenant URL should look something like this "http://core.collectionspace.org:8180/collectionspace/tenant/core/authorities/initialise".  Notice the word "core" in the URL.  And the username for this tenant would look something like this "admin@core.collectionspace.org".

Again, sorry for the trouble and let us know if any of this helps.

-Richard

From: Talk talk-bounces@lists.collectionspace.org on behalf of RAY, Eric Eric.Ray@victoriacollege.edu
Sent: Friday, April 29, 2016 3:02 PM
To: Ray Lee
Cc: talk@lists.collectionspace.org
Subject: Re: [Talk] ant build.xml error on install

This is still happening when I shut down and restart tomcat fresh, and then run the initialization from the script instead of the browser.

I’m sure I’ve done something simple wrong, but I can’t figure out what.

From: Ray Lee [mailto:rhlee@berkeley.edu]
Sent: Friday, April 29, 2016 4:54 PM
To: RAY, Eric Eric.Ray@victoriacollege.edu
Cc: Aron Roberts aron@socrates.berkeley.edu; talk@lists.collectionspace.org
Subject: Re: [Talk] ant build.xml error on install

I've seen the ConcurrentUpdateException happen when two initializations get kicked off simultaneously, which is usually because your browser decides to prefetch the authorities/initialise page as you're typing in the URL. Then you hit enter, and it fetches again while the first fetch is still going. It's CollectionSpace's fault for violating the convention that GET requests are for read (not write) operations. You can diagnose this after the fact by looking in the tomcat access log to see if two GETs were received when you only expected one.

Sometimes two initializations will interleave in such a way that they both succeed, and you end up with duplicated entries.

Usually I end up tailing the cspace-app log while typing in the URL, and if stuff starts happening, I don't hit enter. It's awkward, but at least you don't have to initialize authorities very often.

Ray

On Fri, Apr 29, 2016 at 2:20 PM, RAY, Eric <Eric.Ray@victoriacollege.edumailto:Eric.Ray@victoriacollege.edu> wrote:

More issues! I cloned the core tenant using the ‘easy way’ described in the documentation. That appears to have worked, and I can log in.

When I try to initialize either with the URL or the script, there’s a failure.

Here’s what I have from the browser.

testing Authority  vocab

need to create Instance vocab-xxx

that was weird but probably not a problem Connection exceptionorg.nuxeo.ecm.core.api.ConcurrentUpdateException: Failed to save session: org.nuxeo.ecm.core.api.ConcurrentUpdateDocumentException: org.nuxeo.ecm.core.storage.ConcurrentUpdateStorageException: java.sql.BatchUpdateException: Batch entry 0 INSERT INTO "vocabularies_common" ("id", "vocabtype", "shortidentifier", "description", "source", "displayname", "refname") VALUES ('ffacf4c5-0e0a-4c8a-8f1a-85bb2c3e7c27', NULL, 'xxx', NULL, NULL, 'Temporary Vocab whilst none exist', 'urn:cspace:museumofthecoastalbend.org:vocabularies:name(xxx)''Temporary Vocab whilst none exist''') was aborted.  Call getNextException to see the cause.

I tried cloning a new tenant and trying again – same issue.

From: aronroberts@gmail.commailto:aronroberts@gmail.com [mailto:aronroberts@gmail.commailto:aronroberts@gmail.com] On Behalf Of Aron Roberts
Sent: Friday, April 29, 2016 3:01 PM

To: RAY, Eric <Eric.Ray@victoriacollege.edumailto:Eric.Ray@victoriacollege.edu>
Cc: Richard Millet <richard.millet@lyrasis.orgmailto:richard.millet@lyrasis.org>; talk@lists.collectionspace.orgmailto:talk@lists.collectionspace.org
Subject: Re: [Talk] ant build.xml error on install

On Fri, Apr 29, 2016 at 12:46 PM, RAY, Eric <Eric.Ray@victoriacollege.edumailto:Eric.Ray@victoriacollege.edu> wrote:

While the csadmin user was set correctly in pg_hba.conf, the localhost entry was set to ident instead of trust.  I don’t know why – must have come default. I changed it to trust, and the build worked. I can now log in with the default credentials.

Onward and upward…

Nice - thanks for reporting on your findings here!

The automated installer didn’t work when I tried it – I’m on CentOS. I can’t remember what the issue was, but I just went ahead with the regular installation process.

Makes sense. The CollectionSpace v4.3 installer has been tested with Ubuntu 14.04 LTS, but not so far with other Linux distributions or releases.

We've done a successful install using the v4.2 installer on CentOS 7, while there's a known issue installing on CentOS 6.x:

https://issues.collectionspace.org/browse/CSPACE-6518

And despite that success, it looks there's still some work remaining to do for the OpenJDK install in CentOS 7:

https://issues.collectionspace.org/browse/CSPACE-6721

I'm hoping at some point that installer can be regularly updated to work with (at least) the then-current, long-term support releases of both Ubuntu and CentOS.

Thanks,

Aron

From: aronroberts@gmail.commailto:aronroberts@gmail.com [mailto:aronroberts@gmail.commailto:aronroberts@gmail.com] On Behalf Of Aron Roberts
Sent: Friday, April 29, 2016 1:41 PM

To: RAY, Eric <Eric.Ray@victoriacollege.edumailto:Eric.Ray@victoriacollege.edu>
Cc: Richard Millet <richard.millet@lyrasis.orgmailto:richard.millet@lyrasis.org>; talk@lists.collectionspace.orgmailto:talk@lists.collectionspace.org
Subject: Re: [Talk] ant build.xml error on install

Hi Eric,

This could potentially mean that, in PostgreSQL's host-based authentication file, pg_hba.conf, the authentication method for the 'csadmin' user is set to 'ident', which would require that database access requests come from an operating system user with that same name as that PostgreSQL user ('csadmin'). To resolve this, the relevant entry in that configuration file should likely be changed from 'ident' to 'md5', to allow (hashed) password-based authentication.

For instructions, including an example pg_hba.conf file, see the "Changes to pg_hba.conf" section in this document:
https://wiki.collectionspace.org/display/DOC/PostgreSQL+Installation+under+Linux#PostgreSQLInstallationunderLinux-Changestopg_hba.conf

And as background, there's a description of PostgreSQL's various authentication methods here:
http://www.postgresql.org/docs/current/static/auth-pg-hba-conf.html

BTW, have you had a chance yet to try out the automated CSpace installer? We'd love to have your feedback on it:
https://wiki.collectionspace.org/display/DOC/Automated+installer+for+CollectionSpace+4.3

The goal of that installer is to allow entering a couple of commands, then walk away for about 30 minutes, and come back to a (nearly) complete installation, with just a few finishing steps to perform. There are a lot of steps in a CSpace installation, and we'd like for this to be far less painful.

Thanks,

Aron

On Fri, Apr 29, 2016 at 11:11 AM, RAY, Eric <Eric.Ray@victoriacollege.edumailto:Eric.Ray@victoriacollege.edu> wrote:

As it happens, I do have another issue that’s just come up.

I had gotten the login page running, but realized that logins (even the default core login) were failing, and perhaps that had to do with outdated postgres. So I upgraded postgres to 9.5, erased the tomcat and collectionspace-source, and started over. All went fine, except that in the ant undeploy deploy create_db -Drecreate_db=true import command, I’m getting the following error:

BUILD FAILED

/home/mcb/collectionspace-source/services/build.xml:220: The following error occurred while executing this line:

/home/mcb/collectionspace-source/services/build.xml:256: org.postgresql.util.PSQLException: FATAL: Ident authentication failed for user "csadmin"

I believe that csadmin is set up in postgres correctly, and I believe that all my environment variables are set. I can echo the environment variables for passwords, and they match what I set the csadmin password to be.

Any thoughts?

From: aronroberts@gmail.commailto:aronroberts@gmail.com [mailto:aronroberts@gmail.commailto:aronroberts@gmail.com] On Behalf Of Aron Roberts
Sent: Friday, April 29, 2016 1:03 PM
To: RAY, Eric <Eric.Ray@victoriacollege.edumailto:Eric.Ray@victoriacollege.edu>
Cc: Richard Millet <richard.millet@lyrasis.orgmailto:richard.millet@lyrasis.org>; talk@lists.collectionspace.orgmailto:talk@lists.collectionspace.org
Subject: Re: [Talk] ant build.xml error on install

Thanks much, Eric, for this update. Glad that you were able to get past this stuck point! Appreciate your trying out CollectionSpace, and please don't hesitate to ask any other questions that come up during the process.

In addition to what Richard mentioned, one other possibility that came to mind was file access permissions, for various files/directories in the Tomcat server folder. I've run into those issues when expanding that folder tarball and/or when running some install commands, as root (using 'sudo'), and then running some later commands as an unprivileged user. That might not have been applicable in this case, but am mentioning this in case it helps anyone going forward ...

Aron Roberts

UC Berkeley

On Fri, Apr 29, 2016 at 6:44 AM, RAY, Eric <Eric.Ray@victoriacollege.edumailto:Eric.Ray@victoriacollege.edu> wrote:

This error, for future reference, must have come from an error of mine in previous steps. I started over and it wasn't present the second time around.

Thanks!

On Apr 28, 2016, at 6:35 PM, Richard Millet <richard.millet@lyrasis.orgmailto:richard.millet@lyrasis.org> wrote:

Ray,

Just a guess but this error message seems to indicate that the file (or its parent directory) is read-only or perhaps in-use?  Maybe the tomcat server is still running?

-Richard

From: Talk <talk-bounces@lists.collectionspace.orgmailto:talk-bounces@lists.collectionspace.org> on behalf of RAY, Eric <Eric.Ray@victoriacollege.edumailto:Eric.Ray@victoriacollege.edu>
Sent: Thursday, April 28, 2016 4:19 PM
To: talk@lists.collectionspace.orgmailto:talk@lists.collectionspace.org
Subject: [Talk] ant build.xml error on install

Hello,

I’m working my way through a CollectionSpace install.

When I run the ant undeploy deploy create_db -Drecreate_db=true import command, I get the following errors. At first ant was out of date, but I updated it to 1.9.7, and this is verified by ant –version.

BUILD FAILED

/home/mcb/collectionspace-source/services/build.xml:313: The following error occurred while executing this line:

/home/mcb/collectionspace-source/services/services/build.xml:235: The following error occurred while executing this line:

/home/mcb/collectionspace-source/services/services/common/build.xml:183: The following error occurred while executing this line:

/home/mcb/collectionspace-source/services/services/common/build.xml:199: Unable to delete file /usr/local/share/apache-tomcat-7.0.57/lib/spring-security-acl-3.0.5.RELEASE.jar

Any thoughts?

Eric Ray

Curator of Collections and Exhibits

Museum of the Coastal Bend at Victoria College

2200 E. Red River, Victoria, TX 77901

361-572-6468tel:361-572-6468 | eric.ray@victoriacollege.edumailto:eric.ray@victoriacollege.edu

Talk mailing list
Talk@lists.collectionspace.orgmailto:Talk@lists.collectionspace.org
http://lists.collectionspace.org/mailman/listinfo/talk_lists.collectionspace.org

Talk mailing list
Talk@lists.collectionspace.orgmailto:Talk@lists.collectionspace.org
http://lists.collectionspace.org/mailman/listinfo/talk_lists.collectionspace.org

This worked. I'm working now to configure cspace for our museum - thanks so much for your help.

From: Richard Millet [mailto:richard.millet@lyrasis.org]
Sent: Friday, April 29, 2016 7:42 PM
To: RAY, Eric Eric.Ray@victoriacollege.edu
Cc: talk@lists.collectionspace.org
Subject: Re: [Talk] ant build.xml error on install

Eric,

I'm really sorry you're still having trouble with the install.  Here are a couple other things to try:

1. Shutdown tomcat and manually delete any and all CollectionSpace related databases that were created in your postgres server.  If you haven't changed the CSPACE_INSTANCE environment variable, all the databases created by CollectionSpace should end with "_default".  You can also delete all the associated login roles that also end with "_default".  I use the pgAdmin tool to do these type of tasks, but you can also do this using the psql command line tool.

2. From the "services" layer's root directory, run "mvn clean install -DskipTests"

3. From the "application" layer's root source directory, run "mvn clean install -DskipTests"

4. Back to the "services" layer's root directory, run "ant undeploy deploy create_db -Drecreate_db=true import"

5. Restart tomcat

6. Make sure the URL of the initialization script has the correct tenant ID and that the script has the correct tenant credentials.  For example, the core/default tenant URL should look something like this "http://core.collectionspace.org:8180/collectionspace/tenant/core/authorities/initialise".  Notice the word "core" in the URL.  And the username for this tenant would look something like this "admin@core.collectionspace.orgmailto:admin@core.collectionspace.org".

Again, sorry for the trouble and let us know if any of this helps.

-Richard

From: Talk <talk-bounces@lists.collectionspace.orgmailto:talk-bounces@lists.collectionspace.org> on behalf of RAY, Eric <Eric.Ray@victoriacollege.edumailto:Eric.Ray@victoriacollege.edu>
Sent: Friday, April 29, 2016 3:02 PM
To: Ray Lee
Cc: talk@lists.collectionspace.orgmailto:talk@lists.collectionspace.org
Subject: Re: [Talk] ant build.xml error on install

This is still happening when I shut down and restart tomcat fresh, and then run the initialization from the script instead of the browser.

I'm sure I've done something simple wrong, but I can't figure out what.

From: Ray Lee [mailto:rhlee@berkeley.edu]
Sent: Friday, April 29, 2016 4:54 PM
To: RAY, Eric <Eric.Ray@victoriacollege.edumailto:Eric.Ray@victoriacollege.edu>
Cc: Aron Roberts <aron@socrates.berkeley.edumailto:aron@socrates.berkeley.edu>; talk@lists.collectionspace.orgmailto:talk@lists.collectionspace.org
Subject: Re: [Talk] ant build.xml error on install

I've seen the ConcurrentUpdateException happen when two initializations get kicked off simultaneously, which is usually because your browser decides to prefetch the authorities/initialise page as you're typing in the URL. Then you hit enter, and it fetches again while the first fetch is still going. It's CollectionSpace's fault for violating the convention that GET requests are for read (not write) operations. You can diagnose this after the fact by looking in the tomcat access log to see if two GETs were received when you only expected one.

Sometimes two initializations will interleave in such a way that they both succeed, and you end up with duplicated entries.

Usually I end up tailing the cspace-app log while typing in the URL, and if stuff starts happening, I don't hit enter. It's awkward, but at least you don't have to initialize authorities very often.

Ray

On Fri, Apr 29, 2016 at 2:20 PM, RAY, Eric <Eric.Ray@victoriacollege.edumailto:Eric.Ray@victoriacollege.edu> wrote:

More issues! I cloned the core tenant using the 'easy way' described in the documentation. That appears to have worked, and I can log in.

When I try to initialize either with the URL or the script, there's a failure.

Here's what I have from the browser.

testing Authority  vocab

need to create Instance vocab-xxx

that was weird but probably not a problem Connection exceptionorg.nuxeo.ecm.core.api.ConcurrentUpdateException: Failed to save session: org.nuxeo.ecm.core.api.ConcurrentUpdateDocumentException: org.nuxeo.ecm.core.storage.ConcurrentUpdateStorageException: java.sql.BatchUpdateException: Batch entry 0 INSERT INTO "vocabularies_common" ("id", "vocabtype", "shortidentifier", "description", "source", "displayname", "refname") VALUES ('ffacf4c5-0e0a-4c8a-8f1a-85bb2c3e7c27', NULL, 'xxx', NULL, NULL, 'Temporary Vocab whilst none exist', 'urn:cspace:museumofthecoastalbend.org:vocabularies:name(xxx)''Temporary Vocab whilst none exist''') was aborted.  Call getNextException to see the cause.

I tried cloning a new tenant and trying again - same issue.

From: aronroberts@gmail.commailto:aronroberts@gmail.com [mailto:aronroberts@gmail.commailto:aronroberts@gmail.com] On Behalf Of Aron Roberts
Sent: Friday, April 29, 2016 3:01 PM

To: RAY, Eric <Eric.Ray@victoriacollege.edumailto:Eric.Ray@victoriacollege.edu>
Cc: Richard Millet <richard.millet@lyrasis.orgmailto:richard.millet@lyrasis.org>; talk@lists.collectionspace.orgmailto:talk@lists.collectionspace.org
Subject: Re: [Talk] ant build.xml error on install

On Fri, Apr 29, 2016 at 12:46 PM, RAY, Eric <Eric.Ray@victoriacollege.edumailto:Eric.Ray@victoriacollege.edu> wrote:

While the csadmin user was set correctly in pg_hba.conf, the localhost entry was set to ident instead of trust.  I don't know why - must have come default. I changed it to trust, and the build worked. I can now log in with the default credentials.

Onward and upward...

Nice - thanks for reporting on your findings here!

The automated installer didn't work when I tried it - I'm on CentOS. I can't remember what the issue was, but I just went ahead with the regular installation process.

Makes sense. The CollectionSpace v4.3 installer has been tested with Ubuntu 14.04 LTS, but not so far with other Linux distributions or releases.

We've done a successful install using the v4.2 installer on CentOS 7, while there's a known issue installing on CentOS 6.x:

https://issues.collectionspace.org/browse/CSPACE-6518

And despite that success, it looks there's still some work remaining to do for the OpenJDK install in CentOS 7:

https://issues.collectionspace.org/browse/CSPACE-6721

I'm hoping at some point that installer can be regularly updated to work with (at least) the then-current, long-term support releases of both Ubuntu and CentOS.

Thanks,

Aron

From: aronroberts@gmail.commailto:aronroberts@gmail.com [mailto:aronroberts@gmail.commailto:aronroberts@gmail.com] On Behalf Of Aron Roberts
Sent: Friday, April 29, 2016 1:41 PM

To: RAY, Eric <Eric.Ray@victoriacollege.edumailto:Eric.Ray@victoriacollege.edu>
Cc: Richard Millet <richard.millet@lyrasis.orgmailto:richard.millet@lyrasis.org>; talk@lists.collectionspace.orgmailto:talk@lists.collectionspace.org
Subject: Re: [Talk] ant build.xml error on install

Hi Eric,

This could potentially mean that, in PostgreSQL's host-based authentication file, pg_hba.conf, the authentication method for the 'csadmin' user is set to 'ident', which would require that database access requests come from an operating system user with that same name as that PostgreSQL user ('csadmin'). To resolve this, the relevant entry in that configuration file should likely be changed from 'ident' to 'md5', to allow (hashed) password-based authentication.

For instructions, including an example pg_hba.conf file, see the "Changes to pg_hba.conf" section in this document:
https://wiki.collectionspace.org/display/DOC/PostgreSQL+Installation+under+Linux#PostgreSQLInstallationunderLinux-Changestopg_hba.conf

And as background, there's a description of PostgreSQL's various authentication methods here:
http://www.postgresql.org/docs/current/static/auth-pg-hba-conf.html

BTW, have you had a chance yet to try out the automated CSpace installer? We'd love to have your feedback on it:
https://wiki.collectionspace.org/display/DOC/Automated+installer+for+CollectionSpace+4.3

The goal of that installer is to allow entering a couple of commands, then walk away for about 30 minutes, and come back to a (nearly) complete installation, with just a few finishing steps to perform. There are a lot of steps in a CSpace installation, and we'd like for this to be far less painful.

Thanks,

Aron

On Fri, Apr 29, 2016 at 11:11 AM, RAY, Eric <Eric.Ray@victoriacollege.edumailto:Eric.Ray@victoriacollege.edu> wrote:

As it happens, I do have another issue that's just come up.

I had gotten the login page running, but realized that logins (even the default core login) were failing, and perhaps that had to do with outdated postgres. So I upgraded postgres to 9.5, erased the tomcat and collectionspace-source, and started over. All went fine, except that in the ant undeploy deploy create_db -Drecreate_db=true import command, I'm getting the following error:

BUILD FAILED

/home/mcb/collectionspace-source/services/build.xml:220: The following error occurred while executing this line:

/home/mcb/collectionspace-source/services/build.xml:256: org.postgresql.util.PSQLException: FATAL: Ident authentication failed for user "csadmin"

I believe that csadmin is set up in postgres correctly, and I believe that all my environment variables are set. I can echo the environment variables for passwords, and they match what I set the csadmin password to be.

Any thoughts?

From: aronroberts@gmail.commailto:aronroberts@gmail.com [mailto:aronroberts@gmail.commailto:aronroberts@gmail.com] On Behalf Of Aron Roberts
Sent: Friday, April 29, 2016 1:03 PM
To: RAY, Eric <Eric.Ray@victoriacollege.edumailto:Eric.Ray@victoriacollege.edu>
Cc: Richard Millet <richard.millet@lyrasis.orgmailto:richard.millet@lyrasis.org>; talk@lists.collectionspace.orgmailto:talk@lists.collectionspace.org
Subject: Re: [Talk] ant build.xml error on install

Thanks much, Eric, for this update. Glad that you were able to get past this stuck point! Appreciate your trying out CollectionSpace, and please don't hesitate to ask any other questions that come up during the process.

In addition to what Richard mentioned, one other possibility that came to mind was file access permissions, for various files/directories in the Tomcat server folder. I've run into those issues when expanding that folder tarball and/or when running some install commands, as root (using 'sudo'), and then running some later commands as an unprivileged user. That might not have been applicable in this case, but am mentioning this in case it helps anyone going forward ...

Aron Roberts

UC Berkeley

On Fri, Apr 29, 2016 at 6:44 AM, RAY, Eric <Eric.Ray@victoriacollege.edumailto:Eric.Ray@victoriacollege.edu> wrote:

This error, for future reference, must have come from an error of mine in previous steps. I started over and it wasn't present the second time around.

Thanks!

On Apr 28, 2016, at 6:35 PM, Richard Millet <richard.millet@lyrasis.orgmailto:richard.millet@lyrasis.org> wrote:

Ray,

Just a guess but this error message seems to indicate that the file (or its parent directory) is read-only or perhaps in-use?  Maybe the tomcat server is still running?

-Richard

From: Talk <talk-bounces@lists.collectionspace.orgmailto:talk-bounces@lists.collectionspace.org> on behalf of RAY, Eric <Eric.Ray@victoriacollege.edumailto:Eric.Ray@victoriacollege.edu>
Sent: Thursday, April 28, 2016 4:19 PM
To: talk@lists.collectionspace.orgmailto:talk@lists.collectionspace.org
Subject: [Talk] ant build.xml error on install

Hello,

I'm working my way through a CollectionSpace install.

When I run the ant undeploy deploy create_db -Drecreate_db=true import command, I get the following errors. At first ant was out of date, but I updated it to 1.9.7, and this is verified by ant -version.

BUILD FAILED

/home/mcb/collectionspace-source/services/build.xml:313: The following error occurred while executing this line:

/home/mcb/collectionspace-source/services/services/build.xml:235: The following error occurred while executing this line:

/home/mcb/collectionspace-source/services/services/common/build.xml:183: The following error occurred while executing this line:

/home/mcb/collectionspace-source/services/services/common/build.xml:199: Unable to delete file /usr/local/share/apache-tomcat-7.0.57/lib/spring-security-acl-3.0.5.RELEASE.jar

Any thoughts?

Eric Ray

Curator of Collections and Exhibits

Museum of the Coastal Bend at Victoria College

2200 E. Red River, Victoria, TX 77901

361-572-6468tel:361-572-6468 | eric.ray@victoriacollege.edumailto:eric.ray@victoriacollege.edu

Talk mailing list
Talk@lists.collectionspace.orgmailto:Talk@lists.collectionspace.org
http://lists.collectionspace.org/mailman/listinfo/talk_lists.collectionspace.org

Talk mailing list
Talk@lists.collectionspace.orgmailto:Talk@lists.collectionspace.org
http://lists.collectionspace.org/mailman/listinfo/talk_lists.collectionspace.org