I performed 4 downloads (the same record for each, with different file type--ALA Legacy/Full Darwin Core/Custom Minimal/Custom Full) this morning, and only two of them generated DOIs or emails. If there were errors with the other two, they should have generated emails, but they didn't.
The two that did complete, and which are also in the list of downloads on doi.ala.org.au are:
The nginx log on prod-bdown-b5 confirms that all of the requests were accepted by biocache-service with HTTP 200 codes, but the biocache-service log file has no reference to the other downloads at all. The second two downloads after the ALA Legacy Format were sent while the first download was executing, but the fourth was triggered 7 seconds after the email was sent for the first download.
Is it possible that there is a hash key being used that eliminates duplicate downloads where the fields are different?
2018-11-16 08:29:59,086 [http-bio-8080-exec-13] DEBUG au.org.ala.biocache.util.QueryFormatUtils (QueryFormatUtils.java:544) - escaping lsid urns urn:catalog:NSW
2018-11-16 08:30:16,023 [biocachedownload-pool-500000-RECORDS_INDEX-0] DEBUG au.org.ala.biocache.service.EmailService (EmailService.java:64) - Send email to : REDACTED@REDACTED
2018-11-16 08:30:16,080 [biocachedownload-pool-500000-RECORDS_INDEX-0] DEBUG au.org.ala.biocache.dao.JsonPersistentQueueDAOImpl (JsonPersistentQueueDAOImpl.java:231) - Removing the download from the queue
2018-11-16 08:30:16,080 [biocachedownload-pool-500000-RECORDS_INDEX-0] INFO au.org.ala.biocache.dao.JsonPersistentQueueDAOImpl (JsonPersistentQueueDAOImpl.java:236) - Deleting /data/cache/downloads/offline1542317337205.json true
2018-11-16 08:30:23,134 [http-bio-8080-exec-4] INFO au.org.ala.biocache.service.AuthService (AuthService.java:189) - authCache requesting: https://auth.ala.org.au/userdetails/userDetails/getUserDetails?userName=REDACTED
I performed 4 downloads (the same record for each, with different file type--ALA Legacy/Full Darwin Core/Custom Minimal/Custom Full) this morning, and only two of them generated DOIs or emails. If there were errors with the other two, they should have generated emails, but they didn't.
The two that did complete, and which are also in the list of downloads on doi.ala.org.au are:
https://doi.org/10.26197/5bede521ba9e1 (ALA Legacy Format)
https://doi.org/10.26197/5bede576371d7 (Custom Full)
The nginx log on prod-bdown-b5 confirms that all of the requests were accepted by biocache-service with HTTP 200 codes, but the biocache-service log file has no reference to the other downloads at all. The second two downloads after the ALA Legacy Format were sent while the first download was executing, but the fourth was triggered 7 seconds after the email was sent for the first download.
Is it possible that there is a hash key being used that eliminates duplicate downloads where the fields are different?