Re: Cache hit percentage
2003-04-11 01:32:06
When using selftunebufpoolsize, there is a maximum amount that the
bufpoolsize will increase to - it will never be more that 10% of the
available real memory. They way you are doing it may be the better way - I
would just caution you to watch for paging on your TSM server.
At 02:31 PM 4/10/2003 -0400, you wrote:
No I don't because I don't think it could help me. It would just increase
the bufpoolsize after every expiration which I already do.
Guillaume Gilbert
Storage Administrater
CGI Montreal
Dave Canan <ddcanan AT ATTGLOBAL DOT NET>@VM.MARIST.EDU> on 2003-04-10 14:29:15
Veuillez répondre à "ADSM: Dist Stor Manager" <ADSM-L AT VM.MARIST DOT EDU>
Envoyé par : "ADSM: Dist Stor Manager" <ADSM-L AT VM.MARIST DOT EDU>
Pour : ADSM-L AT VM.MARIST DOT EDU
cc :
Objet : Re: Cache hit percentage
Are you also using selftunebufpoolsize?
At 02:12 PM 4/10/2003 -0400, you wrote:
>Hi all
>
>I've been living with a cache hit percentage at 97-98% for over 2 years
>now. I've always wanted to get to 99 but have never been able to. I
>finally got my ram doubled
>from 1.5 to 3 gb 2 weeks ago. Since then I've been steadily upping the
>bufpoolsize with no results whatsoever. Here are the details of the system :
>
>F80 with 4 processors and 3 GB of ram
>AIX 5.1
>TSM 4.2.3.3
>Db sits on 15 gb 10k rpm Hitachi disks in a 7700e (raid5). The database is
>alone on the raid group.
>2 FC cards for disks connected through Inrange FC9000.
>
>As you can see the database is not mirrored
>
>Volume Name Copy Volume Name Copy Volume
>Name Copy Available Allocated Free
>(Copy 1) Status (Copy 2) Status (Copy
>3) Status Space Space Space
>
>(MB) (MB) (MB)
>---------------- ------ ---------------- ------ ----------------
>------ --------- --------- --------
>/usr/local/tsm/- Sync'd Undef-
>Undef- 6,800 6,800 0
> bd01/db.bd01 ined ined
>/usr/local/tsm/- Sync'd Undef-
>Undef- 6,800 6,800 0
> bd02/db.bd02 ined ined
>/usr/local/tsm/- Sync'd Undef-
>Undef- 6,800 6,800 0
> bd03/db.bd03 ined ined
>/usr/local/tsm/- Sync'd Undef-
>Undef- 6,800 6,800 0
> bd09/db.bd09 ined ined
>/usr/local/tsm/- Sync'd Undef-
>Undef- 7,168 1,168 6,000
> bd09/db02.bd09 ined ined
>
>My bufpoolsize is now at 921600 (900 mb) and heres the q db f=d :
>
> Available Space (MB): 34,368
> Assigned Capacity (MB): 28,368
> Maximum Extension (MB): 6,000
> Maximum Reduction (MB): 3,512
> Page Size (bytes): 4,096
> Total Usable Pages: 7,262,208
> Used Pages: 5,568,905
> Pct Util: 76.7
> Max. Pct Util: 78.0
> Physical Volumes: 5
> Buffer Pool Pages: 230,400
> Total Buffer Requests: 298,089,579
> Cache Hit Pct.: 97.84
> Cache Wait Pct.: 0.00
> Backup in Progress?: No
> Type of Backup In Progress:
> Incrementals Since Last Full: 0
>Changed Since Last Backup (MB): 32.38
> Percentage Changed: 0.15
>Last Complete Backup Date/Time: 04/10/2003 09:00:56
>
>This is at 2:00pm today. I don't really have a performance problem
>(atleast I hope not...). Expiration runs between 1 and 2 hours examining
>over 4 million files,
>deleting between 100,000 and 500,000.
>
>Heres what vmtune says :
>
>vmtune: current values:
>-p -P -r -R -f -F -N -W
>minperm maxperm minpgahead maxpgahead minfree maxfree pd_npages
>maxrandwrt
>74189 222569 2 8 120 128 65536 128
>
>-M -w -k -c -b -B -u -l -d
>maxpin npswarn npskill numclust numfsbufs hd_pbuf_cnt lvm_bufcnt lrubucket
>defps
>629137 16384 4096 1 186 640 9 131072
1
>
>-s -n -S -L -g -h
>sync_release_ilock nokilluid v_pinshm lgpg_regions lgpg_size
>strict_maxperm
>0 1 0 0 0 0
>
>-t -j -J -z
>maxclient j2_nPagesPer j2_maxRandomWrite j2_nRandomCluster
>222569 32 0 0
>
>-Z -q -Q -y
>j2_nBufferPer j2_minPageReadAhead j2_maxPageReadAhead memory_affinity
>512 2 8 0
>
>-V -i
>num_spec_dataseg spec_dataseg_int
>0 512
>
>PTA balance threshold percentage = 50.0%
>
>number of valid memory pages = 786421 maxperm=30.0% of real memory
>maximum pinable=80.0% of real memory minperm=10.0% of real memory
>number of file memory pages = 343322 numperm=46.3% of real memory
>number of compressed memory pages = 0 compressed=0.0% of real memory
>number of client memory pages = 0 numclient=0.0% of real memory
># of remote pgs sched-pageout = 0 maxclient=30.0% of real memory
>
>I also have a smaller TSM server running on the same machine.The database
>is just under 1 gb and the bufpoolsize is at 256000 and I can't get the
>cache hit percentage
>over 97.
>
>So, are the vmtune settings really that bad? I am not a AIX sysadmin and
>haven't quite understood all this vmtune thingy.
>
>Any help or recommendation will be appreciated.
>
>Thanks in advance
>
>Guillaume Gilbert
>Storage Administrater
>CGI Montreal
Dave Canan
TSM Performance
IBM Advanced Technical Support
ddcanan AT us.ibm DOT com
Dave Canan
TSM Performance
IBM Advanced Technical Support
ddcanan AT us.ibm DOT com
|
|
|