Pool Manager is the process which handles the system buffers shown in "show buffers." If it's trying to allocate 18180, it's probably trying to allocate a huge buffer, which indicates this router is probably terminating some sort of a tunnel, dlsw session, or something like that. The second process is, of course, EIGRP, and it's trying to grab 1500 bytes, which probably means its trying to build a packet for sending out an update, or something along those lines.
The problem isn't that you don't have enough memory to fulfill these requests, it's that the router doesn't have a large enough block of free memory to fulfill this request. The current amount of free memory is 1496216, not much, but it's more than what's being requested, but the largest free block is 79864, which is much smaller. The lowest is 12472. My guess: These attempted allocations occured at the low water mark, when memory was even more fragmented than what you are seeing right now. When you are down close to your lowest, 12472, the largest block size is probably under the 1500 byte mark.
Memory is fragmented moderately bad right now, with only a 79k block largest out of 1.3meg free at the moment.
So.... There could be two problems here. The first is the router doesn't have enougb memory. It looks like you are running an RSM with about 64meg of memory (?), with about 1mb free normally. That's probably not good, I'd want to see more like 6mb free normally. When IOS gets into extremely low memory conditions, it can fragment the memory severely, and never be able to clean the fragmentation up. Since you are seeing moderately bad fragmentation, I'd suspect this as a cause.
The other possibility is that some major network trauma has occured recently, causing memory to drop very low, and fragment heavily. It could be some sort of defect in IOS, since there have been memory fragmentation issues in the past.
My initial prescription, though, for this router, is to take two new memory simms as a start. A possible code upgrade might not be bad, but I would only do that if I were comfortable with some later rev of code.
The *dead* process holds memory that was allocated and owned by some other process which is no longer running. For instance, process A starts, allocates some memory, and passes a reference to that memory to process B. Process A now terminates, leaving the memory "ownerless." When process A stops running, IOS will mark the owner of this memory *dead*. It's important to remember that IOS is a shared memory operating system, all the processes can read and write into the entire memory space (with the exception of some protected memory spaces).
It sounds like you need to pick up a copy of Inside Cisco IOS Software! :-)
In this case, the dead process shows 168908 holding, which isn't too much. It does show quite a bit of free'd, 11195672, which is interesting, and a lot of getbuf's, which means it got buffers from the public buffer pools (those you see when you do a show buffers). It could be that some process was running on your router which ate a lot of memory and buffers, but then was turned off. There's not a lot of ways to track this down, without a lot of knowledge of the actual memory block structure (and even then, it's not really easy, if at all possible).
I would generally rule out *dead* as a part of any problem, unless I could see that some process has been configured then removed, for some reason.
We are pleased to announce availability of Beta software for 16.6.3.
16.6.3 will be the second rebuild on the 16.6 release train targeted
towards Catalyst 9500/9400/9300/3850/3650 switching platforms. We are
looking for early feedback from customers befor...
Introduction Featured Speakers Luis Espejel is the Telecommunications
Manager of IENova, an Oil & Gas company. Currently he works with Cisco
IOS® and Cisco IOS XE platforms, and NX to some extent. He has also
worked as a Senior Engineer with the Routing P...
In this session you can learn more about Layer 3 multicast and the best
practices to identify possible threats and take security measures. It
provides an overview of basic multicast, the best security practices for
use of this technology, and recommendati...