Sunday, July 29, 2007
Part 1: who does what?
Last fall, a new application was released. We've been getting hits on the MQ server dead letter queue ever since then. These are usually indicative of serious system problems.
It took me a while to capture one of the bad messages because they would expire so quickly. With the help of a developer, we figured out that the message was too big for the mainframe backend. Who builds a system without edits for this? A vendor, of course.
In May, I opened a trouble ticket for the application area on the problem after a spate of pages (best practice says I have to monitor the DLQ). I am hoping for some kind of quick fix here. The ticket gets assigned to a BA who involves the lead developer who involves the application/system architect. He gives us some background info and outlines three activities. One of them looks likes a quick fix so I create an assignment for my mainframe counterpart. He promptly rejects it saying that the application developers have to be involved. Doh! I forward him the email with the three activities, but he replies that someone must confirm that the message length can be changed on the backend without impact.
By this time, the application/system architect has left on a lengthy vacation. I ask him to check the application after his return. He says it's not his job! Quoi? He tells me that it belongs to the application developer. But this is a mainframe developer not the frontend developer that the BA involved. The mainframe developer spends a day looking at the program and tells me that the message will be arbitrarily truncated, i.e. there is no quick fix. I waited just over three months for this answer, and I don't understand why the BA or the frontend developer didn't just talk to their mainframe counterpart.
D.
It took me a while to capture one of the bad messages because they would expire so quickly. With the help of a developer, we figured out that the message was too big for the mainframe backend. Who builds a system without edits for this? A vendor, of course.
In May, I opened a trouble ticket for the application area on the problem after a spate of pages (best practice says I have to monitor the DLQ). I am hoping for some kind of quick fix here. The ticket gets assigned to a BA who involves the lead developer who involves the application/system architect. He gives us some background info and outlines three activities. One of them looks likes a quick fix so I create an assignment for my mainframe counterpart. He promptly rejects it saying that the application developers have to be involved. Doh! I forward him the email with the three activities, but he replies that someone must confirm that the message length can be changed on the backend without impact.
By this time, the application/system architect has left on a lengthy vacation. I ask him to check the application after his return. He says it's not his job! Quoi? He tells me that it belongs to the application developer. But this is a mainframe developer not the frontend developer that the BA involved. The mainframe developer spends a day looking at the program and tells me that the message will be arbitrarily truncated, i.e. there is no quick fix. I waited just over three months for this answer, and I don't understand why the BA or the frontend developer didn't just talk to their mainframe counterpart.
D.
