Error updating/deleting records

Faiz · Post by **Faiz** » Tue Jun 04, 2002 11:36 am

I am getting some errors while updating or deleting records from table. Insert does not give any error.
1) Corrupt free free-tree page at 0x136F9DDD in KDBF file /usr/local/apache/htdocs/webinator/supdb/doc_category.blb (bad size/len) in the function kdbf_alloc
2) Could not write new page for B-tree /usr/local/apache/htdocs/webinator/supdb/doc_category.blb in the function btgetnewpage
3) Can't create new root page for B-tree /usr/local/apache/htdocs/webinator/supdb/doc_category.blb in the function fbtinsert
Do I have to re-build the whole table? I tried running kdbfchk on this table and it returned an error as well.
Checking file /usr/local/apache/htdocs/webinator/supdb/doc_category.tbl:
Reading free tree and list
0% 50% 100%
000 ABEND: signal 10 (SIGBUS); exiting

Regards,

Post by **Kai** » Tue Jun 04, 2002 1:02 pm

What version of texis is this (complete output of texis -version)? Also, try running kdbfchk on doc_category.blb, not the .tbl.

Faiz · Post by **Faiz** » Tue Jun 04, 2002 1:41 pm

Its TEXIS 4.0. The output of texis -version is
Texis Web Script (Vortex) Copyright (c) 1996-2002 Thunderstone - EPI, Inc.
Commercial Version 4.00.1016093849 of Mar 14, 2002 (sparc-sun-solaris2.6-64)

I ran kdbfchk on doc_category.tbl and not on doc_category.blb. Shall I run it on doc_category.blb?

Post by **mark** » Tue Jun 04, 2002 2:38 pm

Yes. That's what kai said. For future reference run kdbfchk on the file that is reporting an error.

Faiz · Post by **Faiz** » Tue Jun 04, 2002 3:07 pm

kdbfchk on doc_category.blb quit at halfway.
Checking file /usr/local/apache/htdocs/webinator/supdb/doc_category.blb:
Reading free tree and list
0% 50% 100%
-------------------------------------------------------------------------------
*** Free-pages list corrupt ***
Scanning for orphaned free tree pages
0% 50% 100%
##############################000 ABEND: signal 10 (SIGBUS); exiting
There were no inserts/deletes or updates at the time of running kdbfchk. Is the table corrupt?

Post by **Kai** » Tue Jun 04, 2002 4:28 pm

Yes, it appears so due to the "free-pages list corrupt" message, though kdbfchk should have continued. A typical culprit for this is a truncated file due to running out of disk space on the partition, even just briefly, some time in the past. Check for "No space left on device" messages in vortex.log in the past. At this point you'll have to re-generate the table. You can copy what's readable into a new table by something like "create table mynewtable as select * from oldtable".