Hi Nicolas,
Thanks. I will collect the ipcs data which you suggested.
Another question,for the "prstat", I found there are three sybase processes
as below:
# prstat -a
PID USERNAME SIZE RSS STATE PRI NICE TIME CPU PROCESS/NLWP
10032 sybase 7779M 7762M cpu10 20 0 22:20:52 9.7% dataserver/12
10077 sybase 7779M 7762M sleep 50 0 22:38:45 4.4% dataserver/12
10078 sybase 7779M 7762M sleep 52 0 22:29:03 3.9% dataserver/12
...
# ps -eo user,rss,pid,args | grep '^ *sybase' | sort -r -n -k 2,2
sybase 7948280 10078 /usr/sybase/ASE-12_5/bin/dataserver
-ONLINE:2,0,0xffffffffe9032d32, 10000000000
sybase 7948232 10077 /usr/sybase/ASE-12_5/bin/dataserver
-ONLINE:1,0,0xffffffffe9032d32, 10000000000
sybase 7948176 10032 /usr/sybase/ASE-12_5/bin/dataserver
-d/dev/md/rdsk/d71 -e/usr/sybase/ASE-12_5/i
sybase 1336 10036 /usr/sybase/ASE-12_5/bin/backupserver
-e/usr/sybase/ASE-12_5/install/KONDOR_BS.
sybase 608 10035 /bin/sh /usr/sybase/ASE-12_5/install/RUN_KONDOR_BS
sybase 608 10030 /bin/sh /usr/sybase/ASE-12_5/install/RUN_KONDOR
Does mean there are total 3 sysbase instance running(or actually only one
instance) ? and *each *instance used 7779M vm and 7762M physcial RAM ?
Thanks.
Best Regards,
Simon
>
>
>
> On Thu, Nov 26, 2009 at 4:24 AM, Nicolas Dorfsman <n...@woup.net> wrote:
>
>>
>> Le 22 nov. 2009 à 15:35, Simon a écrit :
>>
>> Hi Nicolas,
>>
>> Thank you.
>>
>> I think adding extra swap device would not be helpful in our case,because
>> the customer appliction created too many stack/heaps,these anonymous memory
>> has to use swap device as back store,if the physical RAM really
>> shortage,more swap device only cause page out cost and performance will be
>> degraded accordingly,right ?
>>
>>
>> Right. Except in one case....DISM use of shared memory.
>> Take a look to this command :
>>
>> # ipcs -ZAm | awk 'NR>1&&$16>0{s+=$10} END{print "total=", s/1024}'
>>
>>
>> If there's something different than 0, your Sybase is using DISM segments.
>> With DISM, there's a double 'reservation'....one at creation time, and
>> another at attach time. 'Reservation' doesn't mean 'utilisation', the only
>> way to have sufficient 'rerservation' space is to add swap. Nothing will be
>> swapped out because it's only a reservation. It's just a security to be
>> able to swap out theses segments in case of DR.
>>
>>
>> Nicolas
>>
>>
>>
>>
>>
>> Thanks.
>> Best Regards,
>> Simon
>>
>>
>> On Sat, Nov 21, 2009 at 8:01 PM, Nicolas Dorfsman <n...@woup.net> wrote:
>>
>>>
>>> Hi Simon,
>>>
>>> Le 21 nov. 09 à 07:27, Simon a écrit :
>>>
>>>
>>>> # ipcs -a
>>>> IPC status from <running system> as of Thu Nov 12 12:05:28 HKT 2009
>>>> T ID KEY MODE OWNER GROUP CREATOR
>>>> CGROUP CBYTES QNUM QBYTES LSPID LRPID STIME RTIME CTIME
>>>> Message Queues:
>>>> T ID KEY MODE OWNER GROUP CREATOR
>>>> CGROUP NATTCH SEGSZ CPID LPID ATIME DTIME CTIME
>>>> Shared Memory:
>>>> m 3 0xe9032d40 --rw------- sybase staff sybase
>>>> staff 3 738803712 1314 2125 20:47:22 no-entry 20:47:14
>>>> m 2 0x51 --rw-rw-r-- root root root
>>>> root 1 2000196 2122 8553 15:15:18 15:15:23 13:14:38
>>>> m 1 0x50 --rw-rw-r-- root root root
>>>> root 1 600196 2121 2121 13:14:38 no-entry 13:14:38
>>>> m 0 0xe9032d32 --rw------- sybase staff sybase
>>>> staff 3 7851147264 1314 2125 13:14:40 13:14:40 13:13:42
>>>> T ID KEY MODE OWNER GROUP CREATOR
>>>> CGROUP NSEMS OTIME CTIME
>>>> Semaphores:
>>>> s 1 0x51 --ra-ra-ra- root root root
>>>> root 6 12:05:28 13:14:38
>>>> s 0 0x50 --ra-ra-ra- root root root
>>>> root 6 12:05:28 13:14:38
>>>>
>>>
>>> I'm not sure about Sybase and ISM/DISM, but you may take a look to
>>> ISM/DISM segments by :
>>>
>>> # ipcs -ZAm | awk 'NR>1&&$16>0{s+=$10} END{print "total=", s/1024}'
>>>
>>> DISM segments may suffer of double allocation algorithm. The Workaround
>>> is to add as many disk-swap-area as your SGBD SHM reservation.
>>> I've wrote an article on that (
>>> http://www.solaris-fr.org/home/docs/base/memoire), but in french for
>>> now, sorry.
>>>
>>> Nicolas
>>
>>
>>
>>
>
_______________________________________________
perf-discuss mailing list
perf-discuss@opensolaris.org
