My office recently upgraded to an 8M optical ethernet connection from our old T1... we were given a free 1841 to replace our 2620...
however, it seems that the 1841 stops responding (or rather slows down to slower than a crawl) every 2.5 days or so of continuous running.
I suspect that the 128MB of RAM on this router is the culprit and am thinking of upgrading the RAM to the max of 384MB. however, since i wasn't able to see the logs (the router just was not responsive at all, so i had to restart it), i'm not certain. And considering that our 2620 only had 64MB of RAM...
can any one advise on this? Could RAM make a big difference? or should i just recommend replacing the router with a 2851 router?
Idealy you could use say snmp to gather the stats every say minute (or more often) then you are more likely to get stats close to the crash.
Have you ttried using the console when the router stops responding?
Enable logging to a syslog server and you will be able to see any out of memory messages there.
I would have thought that an 1841 should have been about OK. I can't find at present on CCO the router performance summaries, maybe someone else knows where they are? They have always been hard to turn up but Cisco have hidden them well this time. They were there a few weeks ago for sure and I am convinced that they will still be there. I have unfortunately lost the link.
thanks for this tip... i forgot about IOS version... but since the router is new, i haven't registered it for smartnet and so can't download new versions (or older) of IOS for it yet...
currently, the router runs C1841-IPBASE-M version 12.4(1c)... I guess I better check cisco's site for any advisory on this IOS...
hi Bod43
thanks for this tip... yeah, i've been looking at the mem stats a bit... even using the "sh proc mem" and "sh proc cpu his"...
the cpu history for the past day or so shows hardly any cpu utilization.... memory reports on the other hand is huge... it's like 100 pages long.
seems memory free keeps getting smaller by the minute...
well, when the router stalls, I'm rushed by the others in the office to get it started so never really got the chance to check the console... but i've enabled logging to syslog yesterday... hopefully, i'll get some more info from that soon...
It's strange that our old 2620 seems to be more capable with less memory though...
Maybe you should reboot the router soon. Before it falls over?
This behaviour is anomalous. Routers' memory use normally stabilises and does not increase after a short while.
One 'legitimate' behaviour that I have seen cause this is the failure to specify a default route. The router then does proxy arp (well the one I saw this on was a L2 only switch) for all of the rubbish traffic the arrives over the internet. Memory fills up. Crash - Reboot.
Do you have a default route configured?
You can identify the process that is using the memory.
sh proc mem
the "holding" column is the one you want.
Should be easy enough to see the one that is bigger than 24M - there will only be one. Most likely.
To be honest this is not likely to allow you to fix it but it just might.
Post the results. post sh buff too (while the router still has memory left).
sh mem ! first few lines sh buff ! all
At least you now have a way to determine when the router is at risk of crash - when it has run out of memory - and can do a controlled reboot before this shate is reached.
Oh, the rest of the sh mem output is only of interest to the IOS developers is my guess.
Cabling-Design.com Forums website is not affiliated with any of the manufacturers or service providers discussed here.
All logos and trade names are the property of their respective owners.