i've verified that throttle on cp is disable:
q throttle
I/O throttling is not currently active
Ready; T=0.01/0.01 13:05:55
now the user have 512Mb, but we see the same utilization:
+------------------------------------------------------------------------------+
| z/VM CPU2066 SERIAL 02900A 1984M DATE 11/27/03 START 13:05:53 END
13:06:23 |
| *
|
| <USERID> %CPU %CP %EM ISEC PAG WSS RES UR PGES SHARE VMSIZE
TYP,CHR,STAT |
| INFRAC01 58 1.8 57 657 .00 46K 46K .0 1 100 512M
VUS,DSC,DISP |
| SITACS01 9.4 .06 9.3 3.6 .00 33K 33K .0 981 100 138M
VUS,DSC,DISP |
| CENTINEL 1.1 .03 1.1 1.5 .00 30K 30K .0 1189 100 128M
VUS,IAB,IDLE |
|
|
|<--- DEVICE ---> <----- DEVICE RDEV DATA ------> <-- MEASUREMENT
FACILITY ->|
| *
|
| DEV TYPE VOLSER IOREQST SEC %Q %ER R %LK LNK PA %UT ACC FPT DCT
CN %CN|
|1138 3390 INFR02 6570 219 .00 .00 .00 1 2 10 0 0 0
0 9.3|
|1208 3390 INFR00 6556 218 .00 .00 .00 1 2 10 0 0 0
0 9.2|
|1308 3390 INFR01 6547 218 .00 .00 .00 1 2 10 0 0 0
0 9.2|
|103B 3390 VAR000 65 2 .00 .00 .00 1 2 1.0 4 0 1
3 .70|
|
|
| <----------- CPU STATISTICS --------> <-- VECTOR ---> <STORAGE><XSTORE>
|
| NC %CPU %US %EM %WT %SY %SP XSI %SC NV %VT %OT RSTR %ST PSEC %XS XSEC
TTM|
|-> 1 71 2.1 68 29 .51 0 3172 88 0 0 0 0 48 0 45 0
0.063|
|<-.. 13 .40 13 87 .11 0 500 90 .. 0 0 0 33 0 62 0
0.160|
+-----------------------<-- 05 LOG ACTIONS INDICATED
-->-----------------------+
and log actions:
2) IO RATE EXCEEDED 222 3390-1138 | 0
3) IO RATE EXCEEDED 225 3390-1208 | 0
4) IO RATE EXCEEDED 223 3390-1308 | 0
5) I/O RATE LIMIT EXCEEDED: INFRAC01 672 SEC | 19
on the linux side:
infrac01:~# cat /proc/meminfo
total: used: free: shared: buffers: cached:
Mem: 525975552 206233600 319741952 0 35540992 144084992
Swap: 134815744 0 134815744
MemTotal: 513648 kB
MemFree: 312248 kB
MemShared: 0 kB
Buffers: 34708 kB
Cached: 140708 kB
SwapCached: 0 kB
Active: 10212 kB
Inactive: 169344 kB
HighTotal: 0 kB
HighFree: 0 kB
LowTotal: 513648 kB
LowFree: 312248 kB
SwapTotal: 131656 kB
SwapFree: 131656 kB
the cache is growing fast, but we still have the same utilization, what
could be the cause for linux to only utilize 10% of the devices?
thanks.
On Thu, 27 Nov 2003, Rob van der Heij wrote:
> On Wed, 2003-11-26 at 21:30, Alejandro Leyva Rabinovich wrote:
>
> > i know that the message "I/O RATE LIMIT EXCEEDED: INFRAC01 643 SEC"
> > reffers to a treshold that rtm have, but, does it have any relationship
> > with cp? does z/vm do something when a user exceed the i/o rate limit?
>
> No, the observation that RTM makes does not go back to CP. It is only
> meant to alert (automated) operators who call you in the middle of the
> night about this (or about storage utilization > 100% ...) Other than
> that there is no reason to adjust the threshold.
>
> You could however set a throttle on the device to make CP limit your I/O
> rate on the device if you wish.
>
> PS I guess the rate of 600 I/O's per second is pretty high and I would
> expect very small channel programs. With the lazy write in Linux one
> would expect filling a database to result in long channel programs
> writing a lot of blocks at once. Maybe you already saturated memory and
> bdflushd is running constantly.
>
> Rob
>
--
Alejandro Leyva Rabinovich.
Jefe de la Unidad Departamental de Soporte Ticnico
(Administracisn de Mainframe).
Direccisn General de Informatica.
Secretarma de Finanzas.
Gobierno del Distrito Federal.
