waiting on a buffer
In
onstat -u output the B flag in the first position indicates "waiting on a
buffer".Does this mean "waiting for data to be read from disk into the buffers" versus
"waiting for a free buffer"?
If so, is this typical? What does it mean if the session is constantly waiting
for a buffer?
I am trying to diagnose a slow report on a test system. All seems normal
except for the "B" flag. Yet, there appears to be plenty of free buffers and
the BTR ratio seems OK.
BTW: This report is reading an index not doing a sequential scan.
#onstat -ur|grep 243
c0000000250767c0 B--PR-- 243 sapr3 - c00000000c609318 0 3 842098 2308
c0000000250767c0 B--PR-- 243 sapr3 - c00000000ba46800 0 3 843055 2308
c0000000250767c0 B--PR-- 243 sapr3 - c00000000be53008 0 3 844332 2308
c0000000250767c0 B--PR-- 243 sapr3 - c00000000b877400 0 3 845553 2308
c0000000250767c0 B--PR-- 243 sapr3 - c00000000b861780 0 3 846793 2308
c0000000250767c0 ---PR-- 243 sapr3 - 0 0 3 848010 2308
c0000000250767c0 B--PR-- 243 sapr3 - c00000000af127b0 0 3 849051 2308
c0000000250767c0 B--PR-- 243 sapr3 - c00000000b7cb530 0 3 850165 2308
#BTR.ksh
Informix Dynamic Server Version 7.31.FD2XE -- On-Line -- Up 3 days 12:11:29 --
829512 Kbytes
Profile
dskreads pagreads bufreads Êched dskwrits pagwrits bufwrits Êched
65553038 561164940 779348220 91.59 1144272 4826726 5815514 80.32
isamtot open start read write rewrite delete commit rollbk
470641148 1445626 7992549 426102872 514260 2495600 39388 19975 8364
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 106762.01 23097.90 337 674
bufwaits lokwaits lockreqs deadlks dltouts ckpwaits compress seqscans
3356595 12 177414358 0 0 100 21164 8401
ixda-RA idx-RA da-RA RA-pgsused lchwaits
48710537 1320497 258724 50236229 1676351
----------------------------------------------------------
ReadAhead Utilization (UR): 99.8936 %
Bufwaits Ratio(BR): 0.592012 %
Avg Buffer Turnover Rate (BTR): 33 per hour
#onstat -b|tail -n2
584 modified, 0 resident, 200000 total, 262144 hash buckets, 2048 buffer size
#onstat -R
Informix Dynamic Server Version 7.31.FD2XE -- On-Line -- Up 3 days 12:14:34 --
829512 Kbytes
6 buffer LRU queue pairs priority levels
# f/m pair total % of length LOW MED_LOW MED_HIGH HIGH
0 f 33344 99.3% 33120 0 33120 0 0
1 m 0.7% 224 1 223 0 0
2 f 33317 99.3% 33084 0 33084 0 0
3 m 0.7% 233 3 229 1 0
4 f 33313 99.2% 33054 0 33054 0 0
5 m 0.8% 259 2 257 0 0
6 f 33321 99.1% 33037 0 33037 0 0
7 m 0.9% 284 1 280 3 0
8 F 33312 99.3% 33092 0 33092 0 0
9 m 0.7% 220 1 219 0 0
10 f 33327 99.2% 33054 0 33054 0 0
11 m 0.8% 273 1 271 1 0
1493 dirty, 199934 queued, 200000 total, 262144 hash buckets, 2048 buffer size
start clean at 60% (of pair total) dirty, or 20000 buffs dirty, stop at 50%
58683075 priority downgrades, 0 priority upgrades
----- Original Message -----
From: "CHRIS COMPTON" <chris.compton@phoenix.gov>
To: <ids@iiug.org>
Sent: Thursday, February 13, 2003 3:42 PM
Subject: waiting on a buffer [381]
> In onstat -u output the B flag in the first position indicates "waiting on
a buffer".
> Does this mean "waiting for data to be read from disk into the buffers"
versus "waiting for a free buffer"?
> If so, is this typical? What does it mean if the session is constantly
waiting for a buffer?
>
onstat -g spi
sort by number of spins
go the lru latches come near the top? If so set LRUS to 127.
> I am trying to diagnose a slow report on a test system. All seems normal
except for the "B" flag. Yet, there appears to be plenty of free buffers and
the BTR ratio seems OK.
>
> BTW: This report is reading an index not doing a sequential scan.
>
> #onstat -ur|grep 243
> c0000000250767c0 B--PR-- 243 sapr3 - c00000000c609318 0 3 842098 2308
> c0000000250767c0 B--PR-- 243 sapr3 - c00000000ba46800 0 3 843055 2308
> c0000000250767c0 B--PR-- 243 sapr3 - c00000000be53008 0 3 844332 2308
> c0000000250767c0 B--PR-- 243 sapr3 - c00000000b877400 0 3 845553 2308
> c0000000250767c0 B--PR-- 243 sapr3 - c00000000b861780 0 3 846793 2308
> c0000000250767c0 ---PR-- 243 sapr3 - 0 0 3 848010 2308
> c0000000250767c0 B--PR-- 243 sapr3 - c00000000af127b0 0 3 849051 2308
> c0000000250767c0 B--PR-- 243 sapr3 - c00000000b7cb530 0 3 850165 2308
>
> #BTR.ksh
> Informix Dynamic Server Version 7.31.FD2XE -- On-Line -- Up 3 days
12:11:29 -- 829512 Kbytes
>
> Profile
> dskreads pagreads bufreads Êched dskwrits pagwrits bufwrits Êched
> 65553038 561164940 779348220 91.59 1144272 4826726 5815514 80.32
>
> isamtot open start read write rewrite delete commit
rollbk
> 470641148 1445626 7992549 426102872 514260 2495600 39388 19975
8364
>
> 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 106762.01 23097.90 337 674
>
> bufwaits lokwaits lockreqs deadlks dltouts ckpwaits compress seqscans
> 3356595 12 177414358 0 0 100 21164 8401
>
> ixda-RA idx-RA da-RA RA-pgsused lchwaits
> 48710537 1320497 258724 50236229 1676351
>
> ----------------------------------------------------------
>
> ReadAhead Utilization (UR): 99.8936 %
> Bufwaits Ratio(BR): 0.592012 %
> Avg Buffer Turnover Rate (BTR): 33 per hour
>
> #onstat -b|tail -n2
>
> 584 modified, 0 resident, 200000 total, 262144 hash buckets, 2048 buffer
size
>
> #onstat -R
>
> Informix Dynamic Server Version 7.31.FD2XE -- On-Line -- Up 3 days
12:14:34 -- 829512 Kbytes
>
> 6 buffer LRU queue pairs priority levels
> # f/m pair total % of length LOW MED_LOW MED_HIGH HIGH
> 0 f 33344 99.3% 33120 0 33120 0 0
> 1 m 0.7% 224 1 223 0 0
> 2 f 33317 99.3% 33084 0 33084 0 0
> 3 m 0.7% 233 3 229 1 0
> 4 f 33313 99.2% 33054 0 33054 0 0
> 5 m 0.8% 259 2 257 0 0
> 6 f 33321 99.1% 33037 0 33037 0 0
> 7 m 0.9% 284 1 280 3 0
> 8 F 33312 99.3% 33092 0 33092 0 0
> 9 m 0.7% 220 1 219 0 0
> 10 f 33327 99.2% 33054 0 33054 0 0
> 11 m 0.8% 273 1 271 1 0
> 1493 dirty, 199934 queued, 200000 total, 262144 hash buckets, 2048 buffer
size
> start clean at 60% (of pair total) dirty, or 20000 buffs dirty, stop at
50%
> 58683075 priority downgrades, 0 priority upgrades
>
>
>
>
>
>
>
>