Add this to your properties.xml: <property name="org.apache.manifoldcf.diagnostics" value="DEBUG"/>
This keeps stuff in memory and dumps a lot to the log as well. I'm afraid that groveling through the logs after a failure to confirm it's the same kind of thing we've seen before takes many hours. I can only promise to do this when I have the time. Karl On Tue, Oct 23, 2018 at 2:34 AM Gustavo Beneitez <[email protected]> wrote: > I Karl, > > MySQL. As per config variables: > version 5.7.23-log > version comment MySQL Community Server (GPL) > > which file should I enable logging/debugging? > > Thanks! > > El lun., 22 oct. 2018 a las 21:36, Karl Wright (<[email protected]>) > escribió: > >> Hi Gustavo, >> >> I have seen this error before; it is apparently due to the database >> failing to properly gate transactions and behave according to the >> concurrency model selected for a transaction. We have a debugging setting >> you can configure which logs the needed information so that forensics get >> dumped, and when they do, it's apparent what is happening. >> >> Note well that I have never been able to make this problem appear here, >> so I suspect that the issue is related to network latency or some other >> external factor I cannot easily reproduce. >> >> Just so I know -- what database is this? The place where we've seen this >> is postgresql; later versions of MySql do not seem to have an issue. >> >> Thanks, >> Karl >> >> >> On Mon, Oct 22, 2018 at 1:44 PM Gustavo Beneitez < >> [email protected]> wrote: >> >>> Hi Karl, >>> >>> lately we are facing job status problems. After a few minutes the job >>> ends suddenly, always the same way: >>> >>> Error: Unexpected jobqueue status - record id 1539339908660, expecting >>> active status, saw 2 >>> Error: Unexpected jobqueue status - record id 1539291541171, expecting >>> active status, saw 2 >>> Error: Unexpected jobqueue status - record id 1539294182173, expecting >>> active status, saw 2 >>> Error: Unexpected jobqueue status - record id 1539338949797, expecting >>> active status, saw 2 >>> >>> I did some investigations and a select to the database after the error >>> appeared and BEFORE rerunning the job: >>> >>> SELECT * FROM `jobqueue` WHERE id = 1539336459053 and jobid = >>> 1539269973731 >>> >>> >>> it returned status = 'G' >>> >>> >>> After the run was repeated, it finished OK and same query returned >>> status = 'C'. >>> >>> I don't understand much of the "active" workers but it seems the item is >>> processed twice. Do you have an idea about what we should investigate? >>> >>> >>> Thanks in advance! >>> >>
