Non-reproducible fsync() error


#1

Aha, I think I am getting somehwere. I see this in scidb.log:

2012-10-11 08:41:24,552 [0x7fa492284700] [ERROR]: Error on processing query 1102373729637 on instance 1. Message: SystemException in file: src/smgr/io/Storage.cpp function: flush line: 2016
Error id: scidb::SCIDB_SE_STORAGE::SCIDB_LE_OPERATION_FAILED
Error description: Storage error. Operation 'fsync' failed.
Failed query id: 1102373729637

I did this a few times on my own data, so I thought it was all my fault. Turns out, it’s neither mine or SciDB’s fault. I ran some tests with the Lab exercises with MODIS data. I did the same thing four times in a row, and look!

[code]$ time iquery -nq "LOAD granule2 FROM INSTANCE 0 ‘$DIR/granule2.scidb’ AS ‘OPAQUE’"
SystemException in file: src/smgr/io/Storage.cpp function: flush line: 2016
Error id: scidb::SCIDB_SE_STORAGE::SCIDB_LE_OPERATION_FAILED
Error description: Storage error. Operation ‘fsync’ failed.
Failed query id: 1100873705383

real 0m27.066s
user 0m0.008s
sys 0m0.008s
$ time iquery -nq "LOAD granule2 FROM INSTANCE 0 ‘$DIR/granule2.scidb’ AS ‘OPAQUE’"
SystemException in file: src/smgr/io/Storage.cpp function: flush line: 2016
Error id: scidb::SCIDB_SE_STORAGE::SCIDB_LE_OPERATION_FAILED
Error description: Storage error. Operation ‘fsync’ failed.
Failed query id: 1100883355473

real 0m24.664s
user 0m0.008s
sys 0m0.008s
$ time iquery -nq "LOAD granule2 FROM INSTANCE 0 ‘$DIR/granule2.scidb’ AS ‘OPAQUE’"
Query was executed successfully

real 0m21.851s
user 0m0.008s
sys 0m0.008s
$ time iquery -nq "LOAD granule2 FROM INSTANCE 0 ‘$DIR/granule2.scidb’ AS ‘OPAQUE’"
Query was executed successfully

real 0m23.113s
user 0m0.008s
sys 0m0.008s
$ [/code]
Twice it didn’t work, twice it did. So, this is a real hardware problem, it seems!

I will cut out instance one from the pool and try again.

Cheers, George