iControl failures due to host memory settings?
Hey gang, hoping someone has run into this before. Basically, I have an LTM1500 running 9.4.8 407.0 Hotfix HF4 (the latest in 9.4.8) and for some reason, I have certain iControl calls that fail consistently. For instance, I can do Network/VLAN->get_list and it works as expected, I get a list of VLAN names. When I pass the list of VLAN names to get_vlan_id the server times out and I get "500 server connection timed out before returning any data". This is also true of other iControl methods such as when I call LocalLB/Pool->get_member with a list of pools returned by LocalLB/Pool->get_list.
I can't make heads or tails of this, I have another LTM1500 which works perfectly with the same exact Perl code and is running very old software (9.4.0 Build 517.5 Final).
The only known issue I can find is this one:
Memory use and TMM (CR58673)
Certain iControl operations (such as requesting a list of nodes and then statistics for those nodes) can trigger significant memory pressure, which causes the Traffic Management Microkernel (TMM) to stop running until resources are recaptured. You can work around the problem by increasing the amount of memory allocated to the host. TMM ordinarily leaves 224 MB for the host; increasing memory to 256 MB alleviates the problem.
Side note: this sounds completely backwards to me, why would assigning MORE memory to the host relieve memory pressure on the TMM?
When I look at the host memory it looks like it has plenty:
b memory show all
MEMORY STATISTICS --
| (Host) Total = 748.7MB Used = 733.5MB
| (TMM) Total = 356.0MB Used = 31.81MB
I could assign more memory to the host using the db var (currently set to 0):
bigpipe db Provision.ExtraMB
Provision.extraMB = 0
But... should I really do this? It looks like the TMM isn't really using any memory, so I think I could do this safely.
Anyhow, if anyone has run into this before I would appreciate any insight.