----------------------------------------

From: "Bryant Zimmerman" <[email protected]>

Sent: Thursday, September 29, 2011 1:20 PM

To: "pfSense support and discussion" <[email protected]>

Subject: Re: [pfSense] Carp locking up routers.

----------------------------------------

From: "Chris Buechler" <[email protected]>

Sent: Thursday, September 29, 2011 12:41 PM

To: "pfSense support and discussion" <[email protected]>

Subject: Re: [pfSense] Carp locking up routers.


On Thu, Sep 29, 2011 at 11:06 AM, Bryant Zimmerman <[email protected]> 
wrote:

> I have two identical routers built with supermicro mother boards using 2

> onboard intel nics plus two nic ports on an intel pci-card.

> The issue I am having is that it appears that CARP is locking both units 
up

> at the same time. When the issue occurs I can not reach either router on

> their direct public or private address as soon as I do a hardware rest 
they

> come back to life. When I plugged a console cable in I was able to get

> access to the shell but I did not have the time to do any diag. I just

> issued a reboot command. We are on the nanobsd build. Does anyone have 
any

> ideas. We have seen this on two different production installs. We need 
to

> come up with a solution or we will be unable to use CARP at all.


First guess based on those symptoms (but really guessing, not enough

info to tell you much) - you're running out states, and it's syncing

those to the secondary so both are running out of states.

_______________________________________________

List mailing list

[email protected]

http://lists.pfsense.org/mailman/listinfo/list


Chris 


Thanks for your reply. When the lockup occured we had less then 2MB of 
traffice on the system.

I am unsure if the states would be full but I think it unlikely. I will 
watch them to see.


Current:

State table size


Show states

    MBUF Usage
    11526/25600 


    Based on this should I up my State Table Size?

    My memory never goes above 10% usage either. 


    Thanks Bryant


    Ok I have some more info after looking at the RRD graphs. I noticed a 
major increase in memory usage leading up to the two lockups on one of the 
carp clusters.  Normal memory is at 15% usage. The memory usage increased 
up to 40% before the lockups. Because I am using the nanobsd build I loose 
the last  12 hours of RRD so it does not contain the acutall lockup usages. 
but I can see the increase over approx the 24 hour period before. So there 
is somthing going on there.  Are there logs stored some where that might 
show what is taking the memory over the time covered in the increased 
memory usage times? 


    If I change one of the routers over to a Hard Drive install would I get 
more up to date logs in the event of a reset?


    Thanks Bryant
_______________________________________________
List mailing list
[email protected]
http://lists.pfsense.org/mailman/listinfo/list

Reply via email to