[teiid-issues] [JBoss JIRA] (TEIID-4768) Clarify Caching guide using external materializaion

[ https://issues.jboss.org/browse/TEIID-4768?page=com.atlassian.jira.plugin... ] Ramesh Reddy commented on TEIID-4768: ------------------------------------- There is no 5 minute minimum. When there is error with either source databases or target materialization database (as in they are down and not available), the retry logic has time calculation like below to schedule the next run {code} MATH.min(ttl/4, 60000) {code} So, if TTL was 30 secs, then every 7.5 seconds it will retry. Each time it retries it sets the {{Status}} table to "LOADING", then when it fails then it flips back to "FAILED_LOAD". Now depending upon, if view tables's source database is down, then materialization process fails immediately, that means the {{Status}} could see "FAILED_LOAD" for 7.5 secs before flipped back to "LOADING". If target database is down, then process fails when the materialization process starts writing results to the target database. That means the "LOADING" will appear for amount of time look to write first row to the target database, which can be long time based on view, then when it fails, it flips back to "FAILED_LOAD" for 7.5 seconds before retrying. For user who is monitoring the {{Status}} table manually, this could appear that the {{Status}} table is struck at "LOADING", but in fact that is not the case. See more investigation in TEIID-4758. My suggestion to [~van.halbert] was to use a large enough TTL where he has time to manually verify. -- This message was sent by Atlassian JIRA (v7.2.3#72005)