> From: tech-boun...@lists.lopsa.org [mailto:tech-boun...@lists.lopsa.org] > On Behalf Of Robinson, Greg > > Got an interesting one here. Every random period of time, our Solaris 10 > samba server will stop serving new requests (to map a network drive). > Existing connections still seem to work. At the same time, NETLOGON stops > working as well, and login scripts on windows boxes fail at a random line > when you login. The workaround at the moment is to restart the server > service on the domain controller. This, I believe, will also restart the > NETLOGON service, and all is well.
Did I read that right? Is the samba server acting as a domain controller, not just a file/print server? (see below)... > Our samba server is now pointing to a specific domain controller. If I change > the config to point to the other domain controller, the problem follows the > samba config and that domain controller server service needs to be > restarted. It seems, I did read that right. You have a samba file server, authenticating against an AD domain, and the service dies on the AD server, not the solaris/samba server... You said it's samba 3.5.8, built from source.... So the first places to start are to get the latest version... Validate that you're following the same build process as on other machines... > 2003. There is a project to upgrade them to 2008, but that is taking it's > time. > And we're not even sure that it will fix it. Most likely the upgrade to 2008 will make it worse. I don't know exact details, but they cutover from default NTLMv1 with support for v2... To default v2 with support for v1. Which makes authentication and other activity much slower and prone to failure... > I think it's a rogue box out there, which is hitting the samba server quite > hard, > and the samba server is passing the info onto the DC. I've upped the logging If it's just a load issue, you should be able to see that easily enough by just looking at the samba logs. Most likely it's not a load issue, but some kind of invalid or malformed traffic, which the server can't handle for some reason, so it dies. > If it is a rouge server, how could we find it? If it is malformed traffic, you might have luck simply looking for the latest connection info in samba logs, while in the midst of your next failure... _______________________________________________ Tech mailing list Tech@lists.lopsa.org https://lists.lopsa.org/cgi-bin/mailman/listinfo/tech This list provided by the League of Professional System Administrators http://lopsa.org/
