Incident #9008
closed
algorithmsBot/runAlgorithmsBot.sh going berserk
100%
Description
The algorithmsBot/runAlgorithmsBot.sh script generated hundreds of thousands exceptions in the workspace app. The exceptions are all the same:
22.06.2017 19:22:25.069 [ajp-bio-8009-exec-24] ERROR org.gcube.common.homelibrary.jcr.repository.JCRRepository - Error retrieving Root Node null
22.06.2017 19:22:25.070 [ajp-bio-8009-exec-24] ERROR org.apache.jackrabbit.j2ee.workspacemanager.servlets.rest.Delete - Error deleting item: /Home/statistical.manager/Workspace/DataMinerAlgorithms/parallelized_STEP1_VPA_ICCAT_BF
T_E_Retros.jar
org.gcube.common.homelibrary.home.exceptions.InternalErrorException: org.gcube.common.homelibrary.home.exceptions.InternalErrorException: org.gcube.common.homelibrary.model.exceptions.RepositoryException
at org.gcube.common.homelibrary.jcr.JCRHomeManager.getHome(JCRHomeManager.java:118) ~[home-library-jcr-2.9.0-4.5.0-149341.jar:na]
at org.gcube.common.homelibrary.jcr.JCRHomeManager.getHome(JCRHomeManager.java:148) ~[home-library-jcr-2.9.0-4.5.0-149341.jar:na]
The script should be fixed ASAP.
Updated by Gianpaolo Coro about 8 years ago
Hi Andrea, the bot is "mine". It helps us to produce SAI installation strings. Indeed it does something very simple: a cron on the machine executes it and it asks the messaging service if there are new messages. It does not connect to the Workspace. I guess the messaging system uses the WS instead. One solution is to decrease the frequency of the cron.
Updated by Gianpaolo Coro about 8 years ago
- Assignee changed from Nunzio Andrea Galante to Andrea Dell'Amico
Updated by Andrea Dell'Amico about 8 years ago
- Assignee changed from Andrea Dell'Amico to Roberto Cirillo
The problem seems really caused by that algorithm execution, actually (parallelized_STEP1_VPA_ICCAT_BF, or the one that tries to write that file)
Updated by Roberto Cirillo about 8 years ago
- Status changed from New to Closed
The problem was due to a missing resource on gCubeApps scope (DataAccess-HomeLibraryWebapp). For this reason it was not possible to contact the HL webapp via REST in gCubeApps scope and, consequently, there was a failing request of algorithm publication by SAI.
The bot couldn't serve the publication request by SAI, it was gone in looping on this request, causing a lot of exceptions on HomeLibrary webapp side.
The gCoreEndpoint resource has been added to gCubeApps and the problem is solved.
