| Title: | ase |
| Moderator: | SMURF::GROSSO |
| Created: | Thu Jul 29 1993 |
| Last Modified: | Fri Jun 06 1997 |
| Last Successful Update: | Fri Jun 06 1997 |
| Number of topics: | 2114 |
| Total number of notes: | 7347 |
When talking to a customer last week he identified a need for a way
to flag a service as being "non-relocatable" for a period of time. He
runs large batch operations (2-3 days) and would rather that ASE did
NOT fail over his file system due to ANY events. At other times, when
the large batch updates are not happening, he would like the normal ASE
behaviour.
Is there any way to temporarily flag a service as not relocatable? The
only way I could see was to remove the service at the start of the
batch and re-create it at the end of the batch.
Jc
| T.R | Title | User | Personal Name | Date | Lines |
|---|---|---|---|---|---|
| 1914.1 | USCTR1::ASCHER | Dave Ascher | Mon Mar 03 1997 18:21 | 6 | |
you could remove the failover node(s) from the list of nodes in
the environment. This would have the advantage of not stopping
the service when you made the changes.
| |||||
| 1914.2 | Maybe via the existance of a file | NETRIX::"myrdal@zk3.dec.com" | Greory P. Myrdal | Mon Mar 03 1997 19:41 | 16 |
Hi, You could also have your stop script check for the existence of a file. If it exists, have the stop script fail the stop, thus the service will not be relocated. Outside of ASE you could create or remove the file. The down side to this approach is that there are cases (ie. harware failure) that the relocate could still happen. But there are only a few of these and they involve more serious errors. If you have the option to remove the member from the ASE note .1 is a better solution. -- Greg [Posted by WWW Notes gateway] | |||||
| 1914.3 | USCTR1::ASCHER | Dave Ascher | Mon Mar 03 1997 21:04 | 12 | |
You could also have your stop script check for the existence of a
file. If it exists, have the stop script fail the stop, thus
the service will not be relocated. Outside of ASE you could create
or remove the file.
I use this trick all the time - 'DoNothing' and 'DontStart' are
checked for in my scripts. However, in real life with our bigger
systems the hard failovers are the most common - bad cpu, bad
memory. The script never has a chance to execute.
d
| |||||