Expand my Community achievements bar.

Error installing a package

Avatar

Level 4

I'm trying to install a package of content to AEM 6.1, mongo data store, using the package manager interface.

After a period of time, I get an error "undefined" in the UI. There's a list of nodes that have been created before the error in the Activity Log section of the package manager interface. I tail the error log file, and it looks like nothing is happening. (My thought was perhaps the UI reports an error while the server is still processing the package install). I go to CRXDE Lite and delete the nodes that were created. Then I attempt to install the package again, tailing the log the entire time. The same thing happens, except this time when I get the error "undefined", I see this in the log:

25.07.2015 09:51:53.917 *ERROR* [qtp1285307037-81927] org.apache.jackrabbit.vault.fs.io.Importer Error while committing /content/dam/path/image3_mobile.png/jcr:content/renditions/cq5dam.thumbnail.140.100.png: javax.jcr.InvalidItemStateException: OakMerge0002: OakMerge0002: Conflicting concurrent change. Update operation failed: key: 0:/ update {_collisions.r14ec57ba8eb-0-1=CONTAINS_MAP_ENTRY false, _modified=MAX 1437832310, _collisions.r14ec57a1442-0-1=CONTAINS_MAP_ENTRY false, _collisions.r14ec57a82cd-0-1=CONTAINS_MAP_ENTRY false, _revisions.r14ec579efe8-0-1=SET_MAP_ENTRY c-r14ec57bf8bc-8-1, _revisions.r14ec57a23c2-0-1=SET_MAP_ENTRY c-r14ec57bf8bc-a-1, _revisions.r14ec5796f21-0-1=SET_MAP_ENTRY c-r14ec57bf8bc-5-1, _revisions.r14ec579c629-0-1=SET_MAP_ENTRY c-r14ec57bf8bc-7-1, _collisions.r14ec576d2ef-0-1=CONTAINS_MAP_ENTRY false, _collisions.r14ec576d374-0-1=CONTAINS_MAP_ENTRY false, _collisions.r14ec57a23c2-0-1=CONTAINS_MAP_ENTRY false, _collisions.r14ec5796f21-0-1=CONTAINS_MAP_ENTRY false, _collisions.r14ec579c629-0-1=CONTAINS_MAP_ENTRY false, _revisions.r14ec576d374-0-1=SET_MAP_ENTRY c-r14ec57bf8bc-3-1, _revisions.r14ec576d2ef-0-1=SET_MAP_ENTRY c-r14ec57bf8bc-2-1, _revisions.r14ec57ba8eb-0-1=SET_MAP_ENTRY c-r14ec57bf8bc-f-1, _revisions.r14ec57a1442-0-1=SET_MAP_ENTRY c-r14ec57bf8bc-9-1, _collisions.r14ec579efe8-0-1=CONTAINS_MAP_ENTRY false, _revisions.r14ec57a82cd-0-1=SET_MAP_ENTRY c-r14ec57bf8bc-c-1, _revisions.r14ec57a49d9-0-1=SET_MAP_ENTRY c-r14ec57bf8bc-b-1, _revisions.r14ec5798eda-0-1=SET_MAP_ENTRY c-r14ec57bf8bc-6-1, _revisions.r14ec57abdb3-0-1=SET_MAP_ENTRY c-r14ec57bf8bc-d-1, _revisions.r14ec5777350-0-1=SET_MAP_ENTRY c-r14ec57bf8bc-4-1, _revisions.r14ec57bd64b-0-1=SET_MAP_ENTRY c-r14ec57bf8bc-10-1, _revisions.r14ec57b383f-0-1=SET_MAP_ENTRY c-r14ec57bf8bc-e-1, _collisions.r14ec57bec47-0-1=CONTAINS_MAP_ENTRY false, _revisions.r14ec576b828-0-1=SET_MAP_ENTRY c-r14ec57bf8bc-1-1, _collisions.r14ec57abdb3-0-1=CONTAINS_MAP_ENTRY false, _collisions.r14ec5777350-0-1=CONTAINS_MAP_ENTRY false, _collisions.r14ec57b383f-0-1=CONTAINS_MAP_ENTRY false, _collisions.r14ec57bd64b-0-1=CONTAINS_MAP_ENTRY false, _collisions.r14ec576b828-0-1=CONTAINS_MAP_ENTRY false, _revisions.r14ec57bec47-0-1=SET_MAP_ENTRY c-r14ec57bf8bc-11-1, _collisions.r14ec57a49d9-0-1=CONTAINS_MAP_ENTRY false, _collisions.r14ec5798eda-0-1=CONTAINS_MAP_ENTRY false} (retries 5, 37410 ms)

Again, it appears that nothing is happening in the log, so I delete the nodes that were created in CRXDE Lite, and start researching this exception. Then, 14 minutes later, I notice a similar exception for a different path in the error log, followed by about 100 entries like this:

25.07.2015 10:05:51.337 *WARN* [qtp1285307037-81927] org.apache.jackrabbit.vault.fs.io.Importer Error while committing changes. Retrying import from checkpoint at /content/path2. Retries 1/10
25.07.2015 10:05:51.337 *ERROR* [qtp1285307037-81927] org.apache.jackrabbit.vault.fs.io.Importer E /content/path3 (java.lang.IllegalStateException: Parent node not found.)
25.07.2015 10:05:51.338 *WARN* [qtp1285307037-81927] org.apache.jackrabbit.vault.fs.io.Importer Unable to restore order of /content/path3. Node does not exist.

And then a number of entries indicating it was starting to import content into the dam, and run the rendition workflows on it, which continued for about 20 minutes.

1) What does the exception above mean? Did I actually have 2 package installs running simultaneously?

2) Have I likely corrupted my data store?

3) Is there a way to tell if a package install has completed or not? It seems likely that running two package installs simultaneously (because I don't know that one is still running) would create significant risk of concurrent modifications to nodes.

1 Reply

Avatar

Level 10
  1. From stack trace sounds like an environment issue. Like you have setup active-active cluster instance AsyncIndexUpdate supposed to run on the leader instance of the sling toplogy but due to misconfiguration or product bug running on both the active instance.  Please file an support ticket and attach logs from all the instance at the time of package installation.      In short yes there are 2 concurrent thread operating on same node which are those two thread is not clear from stack trace & needs a connect session. So please take official support route
  2. Probability of data store corruption is very less
  3.  In the activity logs you will see A(Add),E(Error),D(Delete).....   Example:-    "E /content/path3"   means package installed but with some errors.