VMware Cloud Community
ryankey
Contributor
Contributor

Memory leak in SCO 5.0.5 causing network to crash under VMWare ESXi 4.0

I am converting a physical SCO Unix OpenServer 5.0.5 system to modern hardware using VMWare ESXi 4.0.

I've successfully imported the server into VMWare, which was a major pain, but after a week I got it to successfully boot, and it was able to accept client connections on the VM. However, after an hour the SCO server loses network connectivity and drops all client connections. A restart will fix the problem, which will again crash in an hour.

If I try to "tcp start" after the network connection is lost I get an error message about not enough stream resources. I've researched this and bumped up my streams in the scoadmin utility, which only delays the crash but does not fix anything.

Netstat -m reveals that there is a memory leak in the class6 space. I've upgraded the AMD NIC driver to the latest one available on SCO's website, this did not resolve the problem. I am stumped right now, does anybody have any tips?

Thanks

Ryan

Reply
0 Kudos
0 Replies