Redimension failed


#1

I wanted to load 11 kinds of data. (each data size is about 600GB). and cluster is 11 machines (4 CPU, 4TB disk, 32 RAM, 2 instances of each machine).
I finished 1-D array data loading.

I was executing the redimension operator as following command.

 iquery -t -anq "insert(redimension(between(Rrs_678_1D,16000000000,17000000000), Rrs_678),Rrs_678);"
 iquery -t -anq "insert(redimension(between(Rrs_678_1D,17000000000,18000000000), Rrs_678),Rrs_678);"
 iquery -t -anq "insert(redimension(between(Rrs_678_1D,18000000000,19000000000), Rrs_678),Rrs_678);"

Rrs_678 schema is “Rrs_678 value:double [row=0:4319,102,0,col=0:8639,102,0,time=2002185:2014121,2048,0]”

and I showed error as following.

SystemException in file: scr/util/DataStore.cpp function: readData line: 268
Error id: scidb::SCIDB_SE_STORAGE::SCIDB_LE_DATASTORE_CHUNK_CORRUPTED
Error description: Storage error. Chunk header corrupted in DataStore with guid '/work/scidb/data/005/1//datastores/113.data' offset 831552'
SystemException in file: scr/array/MemChunk.cpp function: alignIterators line: 2733
Error id: scidb::SCIDB_SE_EXECUTION::SCIDB_LE_OPERATION_FAILED
Error description: Error during query execution. Operation 'setPosition' failed.
SystemException in file: scr/util/DataStore.cpp function: readData line: 268
Error id: scidb::SCIDB_SE_STORAGE::SCIDB_LE_DATASTORE_CHUNK_CORRUPTED
Error description: Storage error. Chunk header corrupted in DataStore with guid '/work/scidb/data/001/1//datastores/113.data' offset 349312'

after this errors, my scidb doesn’t operate… :frowning: . So, I executed initall of scidb.

I experienced this errors at two time. why occurs corrupted chunk?


#2

Hello,

Thank you for telling us about this. We will investigate and respond as soon as possible.
Can you give us your config.ini file? Can you also provide the scidb.log from the coordinator?


#3

what was the outcome of this investigation?
i ended up reinitializing scidb instance.


#4

in my case the chunk corruption happen scidb instance run out of disk.
after disk was added and instance was restarted the corrupted chunk header errors started showing up