On Tue, 4 May 2004, Julian Coleman wrote:
> I'd be interested to see if you have a similar problem.  Can you look at
> the output from `netstat -m` and the mbuf pools from `vmstat -m`?  I see:
>
>   $ uptime
>   11:37AM  up  9:22, 1 user, load averages: 0.24, 0.17, 0.17

# uptime
 9:46AM  up 23:56, 1 user, load averages: 0.18, 0.20, 0.18


It's strange. I've never had such an uptime without some issues. I had to
reboot the machine yesterday (like every day) but didn't change the
applications that are running. The only thing I've changed is the time
when /var/spool/maillog is to be rotated.
Oh, wait, I've just noticed this in my maillog that I cannot explain:

...
May  5 03:32:08 ti postfix/qmgr[13112]: warning: premature end-of-input on
    private/ether socket while reading input attribute name
May  5 03:32:08 ti postfix/qmgr[13112]: warning: private/ether socket:
    malformed response
May  5 03:32:08 ti postfix/qmgr[13112]: warning: transport ether failure
    -- see a previous warning/fatal/panic logfile record for the problem
    description
May  5 03:32:08 ti postfix/master[305]: warning: process
    /usr/pkg/libexec/postfix/smtp pid 15501 exit status 1
...

This problem has disappeared after a reboot now.

>   $ netstat -m
>   2562 mbufs in use:
>           2562 mbufs allocated to packet headers
>   0 calls to protocol drain routines

# netstat -m
7 mbufs in use:
        7 mbufs allocated to packet headers
0 calls to protocol drain routines


>   $ vmstat -m
>   Memory resource pool statistics
>   Name        Size Requests Fail Releases Pgreq Pgrel Npage Hiwat Minpg Maxpg Idle
>     ...
>   mbpl         256     2571    0        0   162     0   162   162     1   inf    1
>   mclpl       2048        5    0        0     7     0     7     7     4   256    4
>     ...
>   In use 2357K, total allocated 2744K; utilization 85.9%

# vmstat -m | tail
vmstat: Kmem statistics are not being gathered by the kernel.
buf8k       8192     8377    0     5427   423    54   369   369     1   inf    0
mbpl         256      161    0        0    12     0    12    12     1   inf    1
mclpl       2048       25    0        0    17     0    17    17     4   128    4
pnbufpl     1024        7    0        0     2     0     2     2     0   inf    0
extent        24  1480206    0  1480203     1     0     1     1     0   inf    0
ncr53c9x_ecb  56   369471    0   369471     1     0     1     1     0   inf    1
scxspl       152   369485    0   369485    10     0    10    10     0   inf   10

In use 62412K, total allocated 67492K; utilization 92.5%

These is the output after the reboot:
# vmstat -m |tail
vmstat: Kmem statistics are not being gathered by the kernel.
buf8k       8192      212    0        0    27     0    27    27     1   inf    0
mbpl         256       31    0        0     3     0     3     3     1   inf    1
mclpl       2048        3    0        0     6     0     6     6     4   128    4
pnbufpl     1024        3    0        0     1     0     1     1     0   inf    0
extent        24    10184    0    10181     1     0     1     1     0   inf    0
ncr53c9x_ecb  56     2516    0     2516     1     0     1     1     0   inf    1
scxspl       152     2530    0     2530     4     0     4     4     0   inf    4

In use 5429K, total allocated 6608K; utilization 82.2%



I will try the kernel patch from Christos the next time the machine
decides to freeze again.

Christian