Project

General

Profile

Task #887

Updated by Aaron Marcuse-Kubitza over 10 years ago

h3. issue 

 * in importing even moderate-sized datasources (eg. NY, SALVIAS) causes all the last full-database import, this caused available disk space errors in 29 of [[VegBIEN_contents#datasources|41 datasources]]: to be used up, and crashes the import: 
 <pre> 
 ssh -t vegbiendev.nceas.ucsb.edu exec sudo su - aaronmk 
 export version=r13016 version=test_import 
 grep --files-with-matches -F "No space left on device" inputs/{.,}*/*/logs/$version.log.sql 
 # and uniqify by datasource the list includes all the datasources in the test run: NY, SALVIAS 
 </pre> 
 * in a test run, importing even moderate-sized datasources (eg. NY, SALVIAS) causes all the available last full-database import, this caused disk space to be used up, and crashes the import: errors in 29 of [[VegBIEN_contents#datasources|41 datasources]]: 
 <pre> 
 ssh -t vegbiendev.nceas.ucsb.edu exec sudo su - aaronmk 
 export version=test_import version=r13016 
 grep --files-with-matches -F "No space left on device" inputs/{.,}*/*/logs/$version.log.sql 
 # the list includes all the datasources in the test run: NY, SALVIAS and uniqify by datasource 
 </pre> 
 

 * because the test run crashed, as well, the problem this is most more likely a Postgres bug, but could also be a bug in Postgres or Linux itself. this unfortunately means that we *can't import datasources individually, either*, until we find any other part of the Postgres/Linux bug that is causing the problem. 

 VM 
 * this is not necessarily caused by #884, because sort temp files consume disk space only in proportion to the table size, which for a small import would not be nearly enough to fill the disk (and yet the disk does fill up) 

 h3. fix 

 * roll back Postgres to the version it was at in the last successful import, re-run import, and see if problem goes away 
 ** this may require building Postgres from source, because past _revisions_ of the same numeric version might only be available in version control, not in binary form via apt-get (which numbers packages by numeric version) 
 * if this isn't possible, it may be necessary to downgrade to Postgres 9.2 (which will unfortunately be missing some features that we now use)

Back