Joined: 26 May 2005 Posts: 178 Location: Copenhagen, Denmark
I have to send 12 million rows from Z/OS system to another Z/OS Host.
It is a monthly job and on and average the data would be 12M.
I'm wondering whether to FTP the data file or asking them to use the cross loader. They are okay with both the options. Now the option is left to me. I was thinking in terms of cost effective. I have no idea about the FTP concept. Do you guys agree that CROSS loader would be better..?
Instead of Unload,FTP and the LOAD. Please let me know your thoughts.
Joined: 05 Feb 2007 Posts: 199 Location: Sri Lanka
I have never used a Cross Loader and V8 DB2, but from what I understand about how it works, it seems if the job fails due to any reason during load(say after loading 11M rows), you have to start from the beginning again by reading those 12M rows from source. This second-time read and transfer to destination will not be needed if you use a unload/load process because the unloaded rec file will already be there at destination.
And I think, since you need DDF and a three part name for the table to make Cross Loader work anyway, you can probably unload the rows into destination system by running a job from destination system using the typical DSNTIAUL but with a three part name to access the source table at the source system. e.g. select * from ssid.schema.tablename
That way like CrossLoader, you don't have to worry about ftp, but, unlike CrossLoader, time consuming repeat-reads if load fails will be avoided.
Joined: 01 Jul 2003 Posts: 15 Location: Phoenix , AZ
I have used both Cross loader , FTP as well as other IBM replication tools.
I have a small contradiction with what wanderer says , if you unload the data with ssid.schema.tablename and load it onto target table. To my understanding ,
here is it how , it works it first establishes connectivity to the source server thru DDF and unloads the data using the cursor. Once the data is completely unloaded , then the connection is dropped and starts loading at the target tables.
Hence , as Wanderer says , it will restart even , if your load fails.
I would suggest you as follows
1) If this is going to be a one time activity , you can think of getting the data unloaded into neither an unload or an IXF ( which would be comparatively lesser in size ) and then FTP it to the destination and load it to the target table.
2) If this is going to be a repetitive in nature , you can think of replication tools such as IBM Data Propagator or IBM Replidata , which has more feature and gives us better control over replicating the data from one server to another.
Joined: 16 Apr 2007 Posts: 3 Location: Bangalore, INDIA
Did you mean to say Restart is possible in case of cross loader without having to getting connected to source again..? Sorry if i took your point in other way. As you told the data will be unloaded into the cursor but when the job gets failed this will be erased ..right..? As this will be in temporary storage..!