Performance Issues
Hi,
I'll appreciate any thoughts on this matter.
Im currently running an Informix based ( Informix 7.31 FC 5 Database Engine,
D4GL ver 3.0) Database Application on a DEC Alpha ES40 Unix box running
TruUnix 4.0 F. I have roughly 20 disks on a RAID Array system which
facilitate RAID 1 mirroring as well as Informix software mirroring for my
root dbspace. During recent times, I've experienced siginificant performance
degradation on the appliction response time yet when I run some statistics
both on the Indformix side and Unix operating system side, I see no major
violators.
I ran an oncheck command to see whether my database was fragmented; but most
of my tables had only one extent.
When I did an onstat -p,my read and write cache rate were over 99% and 96%
respectively as shown below;
Informix Dynamic Server Version 7.31.FC5 -- On-Line -- Up 26 days
21:28:25 --
352256 Kbytes
Profile
dskreads pagreads bufreads %cached dskwrits pagwrits bufwrits %cached
34133868 4893915 986627967 96.54 806353 2995922 221839041 99.64
isamtot open start read write rewrite delete commit
rollbk
423605965 3461139 18012746 214076922 107648943 496825 85792 107395
47
gp_read gp_write gp_rewrt gp_del gp_alloc gp_free gp_curs
0 0 0 0 0 0 0
ovlock ovuserthread ovbuff usercpu syscpu numckpts flushes
0 0 0 13559.67 1203.08 694 2583
bufwaits lokwaits lockreqs deadlks dltouts ckpwaits compress seqscans
2489966 0 209827275 0 0 71 61277 140953
ixda-RA idx-RA da-RA RA-pgsused lchwaits
1220643 75817 33152539 33275723 12988
In running some check on my o/s, the idle time on my CPU was 99%. There was
no major swapping or paging taking place. and extent of pageouts was
tolerable.
VMSTAT:
Virtual Memory Statistics: (pagesize = 8192)
procs memory pages intr cpu
r w u act free wire fault cow zero react pin pout in sy cs us sy
id
4 229 31 28K 1313 32K 65M 5M 8M 687K 45M 80K 8 230 400 0 0
99
4 234 31 28K 1133 33K 2314 276 406 0 1436 0 23 661 417 0 2
97
4 238 31 28K 971 33K 397 19 219 0 145 0 18 654 393 0 2
98
4 239 31 28K 959 33K 2415 286 410 0 1512 0 23 170 433 0 1
99
4 239 31 28K 959 33K 200 0 200 0 0 0 2 118 320 0 1
99
4 239 31 28K 959 33K 200 0 200 0 0 0 4 147 333 0 1
99
4 239 31 28K 959 33K 200 0 200 0 0 0 1 139 323 0 1
99
4 239 31 28K 947 33K 804 44 252 0 440 0 1 360 358 0 1
99
IOSTAT:
tin tout bps tps bps tps bps tps bps tps us ni sy id
0 12 0 0 5 0 3 0 3 0 0 0 0 99
0 30 0 0 0 0 0 0 0 0 0 0 0100
0 30 0 0 0 0 0 0 0 0 0 0 0100
0 30 0 0 0 0 0 0 0 0 0 0 0100
0 30 0 0 0 0 0 0 0 0 0 0 1 99
0 30 0 0 0 0 0 0 0 0 0 0 0100
Informix Dynamic Server Version 7.31.FC5 -- On-Line -- Up 23 days
18:47:24 --
352256 Kbytes
Segment Summary:
id key addr size ovhd
class blkused blkfree
1 1381451777 200000000 167772160 3456 R* 19542
938
2 1381451778 20a000000 192937984 3448 V 4273
19279
3 1381451779 215800000 8388608 632 M 43
981
Total: - - 369098752 -
- 23858 21198
(* segment locked in memory)
I've been doing an "update statistics" every day and have some experienced
some performance improvement intermittedly.
This machine is 5+ years old and I'm now wondering if the hardware just
doesn't have the capaciity again
I haven't seen anything on the database side that could be affecting
performance; but having my cpu idle 99% of the time also reveals very little
activity on the operating system.
Any thoughts here??
thanks in advance
The information in this communication and any attachments are deemed by PPSL
to be confidential and may be legally privileged and protected from
disclosure. It is intended solely for the addressee. If you are not the
intended recipient, any use, review, dissemination, distribution or copying
of this information is strictly prohibited. Please notify the sender and
delete this message and any attachments immediately from your system.
PPSL will not be held liable for any incomplete transmission of this message
nor any delay in its receipt.
sending to informix-list