Cause for cyclic lock in Sybase 12.0 -
12-04-2003
, 12:31 AM
Hi,
I have a scenario as below.
There is a process A which is related to drop index on table1
There is a process B which is related to bcp on table2
There is a process C which is related to create index on table3
There other processes D, E, F which are also related to tables table4,
table5 and table 6 respectively.
All the above processes are forked parallely from a UNIX shellscript.
All the tables are in the same database.
Now my problem is
process A is blocking process B
process B is blocking process C
process C is blocking process A
process A also blocks process D, E and F.
So, to me it looks like it is entered into a deadlock situation.
I would like to know, what is the reason for this problem. Why have
all the tables in the same database entered into such a situation.
Actually we have implemented in such a way that each process tries to
work on different tables though executed as parallel processes.
The sequence of step done by the shell script on all the tables is
1. BCP data to the table using bcp utility.
2. Create clustered index on the table.
3. Process the data into other tables .
4. truncate the data from table.
5. Drop the index on the table.
The above is done ones' every 15 minutes.
Early help is very much appreciated.
Thanks in advance |