Hi everyone,
We've hit a snag while reindexing one of our tables. Just started yesterday, the table's been around for weeks with no problems before this, and a kdbfchk of the .tbl and .blb show no problems.
When we run these index statements we get the errors described below them:
tsql -q "set keepnoise='on';set delexp=0;set addexp='\punct{1,5}';set addexp='\alnum{1,99}';set addexp='>>\alpha{1,50},=\alpha{1,50}';create metamorph inverted index myindex on mytable(FIELD01\FIELD02\FIELD03.......);"
000 tsql ABEND: signal 11
tsql -q "create unique index myindex_IDFIELD on mytable(IDFIELD):;
002 Can't open KDBF file /home/database01/575/mytable/myindex_D.btr for writing: error 2: No such file or directory
002 Can't open KDBF file /home/database01/575/mytable/myindex_T.btr for writing: error 2: No such file or directory
The "FIELD01\FIELD02\FIELD03...." bit has been shortened for readability. The actual length of that virtual field is ~1100 characters.
If we trim it down and skip the last few fields, to where it's ~990 characters then the reindex works. The fields we've skipped are regular varchars, with not much data in them and they're not special in any way. We are indexing a blob field, but it's included in the 990 character reindex scenario, and since that worked I don't think it's the problem.
The errors come up after about 10 minutes, so they're not immediate. The table itself, .tbl and .blb combined is maybe 1 GB in size-- nothing we haven't seen hundreds of times. We've tried reindexing on several boxes and the same thing happens on each.
I'm thinking there's some funky data (maybe) in our latest table update, but nothing jumps out from previous updates. I've tried playing with the indexmem statement, bumping it down to as low as 100 MB, but all that did was delay the inevitable.
I'm open to suggestions on what we can try next? We're on version 4.04.1067366033 of texis, Linux flavor.
Thanks!
We've hit a snag while reindexing one of our tables. Just started yesterday, the table's been around for weeks with no problems before this, and a kdbfchk of the .tbl and .blb show no problems.
When we run these index statements we get the errors described below them:
tsql -q "set keepnoise='on';set delexp=0;set addexp='\punct{1,5}';set addexp='\alnum{1,99}';set addexp='>>\alpha{1,50},=\alpha{1,50}';create metamorph inverted index myindex on mytable(FIELD01\FIELD02\FIELD03.......);"
000 tsql ABEND: signal 11
tsql -q "create unique index myindex_IDFIELD on mytable(IDFIELD):;
002 Can't open KDBF file /home/database01/575/mytable/myindex_D.btr for writing: error 2: No such file or directory
002 Can't open KDBF file /home/database01/575/mytable/myindex_T.btr for writing: error 2: No such file or directory
The "FIELD01\FIELD02\FIELD03...." bit has been shortened for readability. The actual length of that virtual field is ~1100 characters.
If we trim it down and skip the last few fields, to where it's ~990 characters then the reindex works. The fields we've skipped are regular varchars, with not much data in them and they're not special in any way. We are indexing a blob field, but it's included in the 990 character reindex scenario, and since that worked I don't think it's the problem.
The errors come up after about 10 minutes, so they're not immediate. The table itself, .tbl and .blb combined is maybe 1 GB in size-- nothing we haven't seen hundreds of times. We've tried reindexing on several boxes and the same thing happens on each.
I'm thinking there's some funky data (maybe) in our latest table update, but nothing jumps out from previous updates. I've tried playing with the indexmem statement, bumping it down to as low as 100 MB, but all that did was delay the inevitable.
I'm open to suggestions on what we can try next? We're on version 4.04.1067366033 of texis, Linux flavor.
Thanks!