TCP Reset Issue: "session killed by session db"

I'm not sure if this is an issue with LTM or APM, but here's the setup.

 

2 x virtual servers:

 

1. virtual server vs-virtualserver-stage
2. virtual server vs-virtualserver-prod

Each have a similar APM policy, that bypasses APM for local subnets and if non-local ip's, then process authentication.

 

Both have persistent 1 hr src ip policies.

 

Now, when accessing stage in internet explorer OR chrome in one tab, and prod in another tab, same browser at first appears normal.

 

After a few seconds of working in either tab, accessing the other tab, and issueing a page refresh ends in a err_connection_reset and page not found. Refreshing again, the site is again reachable.

 

When doing a tcpdump from the root shell of the F5 ltm, the tcpdump will show that the F5 virtual server (whichever one is being accessed) will generate a tcp.rst packet with a cause of "session killed by session db"

 

rst_cause="[0x2772d24:24847] Session killed by session DB"

 

While this capture was being done, one was also being done looking for rst packets from the back end systems of the pool and they're not generating one, so this would appear to be from the F5 itself. I've checked K13223 and this particular cause isn't listed, nor does it show up in any searches that I've yet been able to track down. I have opened a support case on this but was just curious if anyone else has ever encountered this or something similar before.

 

Edit:

 

After some additional checking around, I have heard of something similar from our VM administrator that has two separate farms, and he's seen an issue when one browser tab is open to the older farm and another tab is open to his newer 6.x farm from time to time one of his sites management pages will get reset, but these systems are not on the F5 so I'm starting to wonder if this is actually a browser based issue but it happens in either Chrome or MS IE..