Got a problem with one of our servers, this one is a doozy.
One of our companies has been having issues with dropped calls since we put their server in production around the new year. The server is a dell 2850, * 1.2.4, zaptel 1.2.5, running a TE410P, fedora core 4 (runlevel 3, nmi_watchdog enabled, vga mode=normal), and has the standard settings to help with dropped calls (callprogress=no, busydetect=no, etc).
we have two ‘groups’ on this server - customer service and collections. both have inbound queues, although 99% of customer service’s calls are through the queue, and vice versa with collections.
what is happening is that one, some, or all of customer service’s calls will drop, while nothing happens to collections. this happens several times a day, and is seemingly random. also, the customer service group seems to be plauged with sound quality issues (static is the most common complaint, occasional dropouts in the audio) but the dropped calls and sound quality do not occur with any pattern between them.
there are no load issues with this server, and we have ample bandwidth available. the two groups share the same core network switches, so it’s not a bad switch or bad cabling. we have even gone so far as to run a completely dedicated electrical circuit to the customer service group (serving just six cubes, instead of 18 like usual) but that has done NOTHING. in other words, i’m out of ideas.
in case anyone remembers, this was one of the servers having stability issues, which were fixed when we went to a PCI based NIC and disabled the onboard port. the server was up for 28 days without a hiccup (other than the call drops) and was rebooted yesterday, just for good measure.
i’m up against a wall, out of ideas - if you ask me something, chances are i’ve already tried it, but PLEASE don’t hesitate if you have an idea - i’m getting really desperate here.
thanks in advance, and thanks for looking.