SLD Downtime

[Former Member]

In our prod, SLD in local to XI or in other words SLD using XI java engine.

I installed the De-Central Adapter Engine and some of our important interfaces goes through it.

If XI java goes out of memory, it is going to be restarted by its own, till that XI java comes up. SLD will be down

because it uses XI java engine. It takes usually 6 minutes to come up. In mean time will effect my interfaces running through

De-central adapter engine? my understanding for every 3 to 5 minutes XI runs reports to update the cache ? in this time it communicates with SLD and not for every single interface execution.

Can my De-Central adapter will able to manage this SLD downtime ?? If not sure ? is there any way to avoid this ?? by installing another SLD and making a bridge between this SLD and that SLD ?? or like that ?? I'm not sure if this possible ?

Any thoughts and knowledge share will be greatly appreciated.

Thanks

Sagar

[Former Member · Accepted solution]

Hi my friend

Most importantly when SLD is down the integration engine will go down too, so I would suggest we should resolve this issue from the root, which is AS Java server node(s) get restarted because of lack of memory like you described.

If it sounds like a plan, then:

How much physical memory does this server have? How much is it allocated to DB? How many server nodes does XI have? Any other instances or application running on the same physical server?

Regards,

[Former Member · Accepted solution]

Thanks for the response

Hardware is good (3 entitled, 6 virtual P6 CPU's) , memory is 8GB physical free memory is there, 16GB page space is there, CPU utilization I just checked it around 10% of entitled. We've only One server node (Server0). I can easily add 2 more nodes.

But my question here is, lets supposes if I've 2 nodes, one node got OOM (OutOfMemory) its restarts that server node by own or it restarts entire JAVA engine?? (Which means 2 nodes) ??

If it restarts only the node which is going OOM means , JAVA will not go down as one node still up and we may not see SLD down.

Thanks

Sagar

[Former Member · Accepted solution]

Is there any way we can back up the SLD ? if XI java goes down, it should point to another SLD so that XI ABAP (integration engine) will not go down and function accordingly ??

Thanks

Sagar

[Former Member · Accepted solution]

> Is there any way we can back up the SLD ? if XI java goes down, it should point to another SLD so that XI ABAP (integration engine) will not go down and function accordingly ??

Again, you still need to resolve this from the root. Please check my input in this thread:

Additionally, in your case memory allocated for XI ABAP stack also needs to be added up, it's determined by paramter em/initial_size_MB. Overall 8GB = 1GB OS kernel + DB memory + ABAP memory + dispatcher + server node(s), JDK tuning is described in note 723909.