Re: Чего именно не хва тает при zonelimit ?

[Igor Sysoev <is@xxxxxxxxxxxxx>]

On Mon, Dec 22, 2008 at 11:30:57PM +0300, Михаил Монашёв wrote:

> Скажите пожалуйста, как узнать, чего именно не хватает nginx-у, когда
> он находится в состоянии zonelimit?
> 
> Имеется следующая картина:
> 
> last pid: 59174;  load averages: 83.52, 151.66, 178.32 up 24+11:17:27  
> 23:08:30
> 1039 processes:1 running, 1036 sleeping, 2 zombie
> CPU:  2.1% user,  0.0% nice, 27.7% system,  0.0% interrupt, 70.2% idle
> Mem: 1202M Active, 12G Inact, 1954M Wired, 183M Cache, 214M Buf, 421M Free
> Swap: 4096M Total, 1716K Used, 4094M Free
> 
>   PID USERNAME  THR PRI NICE   SIZE    RES STATE  C   TIME   WCPU COMMAND
>  7494 www         1 -16    0 32604K 25184K zoneli 1   0:50  0.00% nginx

> netstat -m
> 47892/77043/124935 mbufs in use (current/cache/total)
> 29891/51363/81254/262144 mbuf clusters in use (current/cache/total/max)
> 29891/3438 mbuf+clusters out of packet secondary zone in use (current/cache)
> 16896/0/16896/16896 4k (page size) jumbo clusters in use 
> (current/cache/total/max)

Вот этого - 4k (page size) jumbo clusters.

> 0/0/0/8448 9k jumbo clusters in use (current/cache/total/max)
> 0/0/0/4224 16k jumbo clusters in use (current/cache/total/max)
> 139339K/121986K/261325K bytes allocated to network (current/cache/total)
> 0/0/0 requests for mbufs denied (mbufs/clusters/mbuf+clusters)
> 0/0/0 requests for jumbo clusters denied (4k/9k/16k)
> 0/0/0 sfbufs in use (current/peak/max)
> 0 requests for sfbufs denied
> 0 requests for sfbufs delayed
> 100053 requests for I/O initiated by sendfile
> 0 calls to protocol drain routines
> 
> 
> vmstat -z
> ITEM                     SIZE     LIMIT      USED      FREE  REQUESTS  
> FAILURES

> mbuf_cluster:            2048,   262144,    33329,    47925, 12408088,        > 0

Увеличивается в sysctl kern.ipc.nmbjumbop

В 7-ке kern.ipc.nmbclusters уже не так критичны - там нужно увеличивать
именно kern.ipc.nmbjumbop.


-- 
Игорь Сысоев
http://sysoev.ru