Solved

Mistuned HP-UX kernel parameters caused occasional surge in Oracle max processes?

Posted on 2009-07-05
9
5,544 Views
Last Modified: 2013-12-19
Hi,

Usually  "select count(*) from v$processes"  during online hours for our PA-RISC
HP-UX B11.11 production server (with 4GB RAM) ranges from 21-80.

We initially run into problems of nfile & nproc being exceeded and we've since
increased these by about 6-7 fold and we don't face this issue anymore.

We have also increased our Oracle processes & sesssions limit from initial 300
to 500 and then to 900.

Things stabilized longer but every 2-3 weeks, the number of Oracle processes
(as given by "select count(*) from v$processes" would exceed 900).  We ran
script to monitor the number of Oracle processes and it rarely go above 100
but when there's a disruption, it just suddenly hit 900.

I can't increase the max processes in Oracle beyond 1200 due a couple of
sem*  kernel parameters.

What could be the cause of the surge in Oracle processes?  Can a mistuned
system cause this or this is mistuned Oracle or Application went wild?

So what should I do next?  There's no error in syslog nor Oracle
listener lsnridb.log logfiles.

In Oracle clients' logs, it says  :
ORA-12500: TNS:listener failed to start a dedicated server process

Our current HP-UX kernel parameters settings are :
# kmtune
Parameter             Current Dyn Planned                    Module     Version
===============================================================================
NSTRBLKSCHED                -  -  2                          
NSTREVENT                  50  -  50                        
NSTRPUSH                   16  -  16                        
NSTRSCHED                   0  -  0                          
STRCTLSZ                 1024  -  1024                      
STRMSGSZ                65535  -  65535                      
acctresume                  4  -  4                          
acctsuspend                 2  -  2                          
aio_listio_max            256  -  256                        
aio_max_ops              2048  -  2048                      
aio_physmem_pct            10  -  10                        
aio_prio_delta_max         20  -  20                        
allocate_fs_swapmap         0  -  0                          
alwaysdump                  1  -  1                          
bcvmap_size_factor          2  -  2                          
bootspinlocks               -  -  256                        
bufcache_hash_locks       128  -  128                        
bufpages                    0  -  (NBUF*2)                  
chanq_hash_locks          256  -  256                        
core_addshmem_read          0  Y  0                          
core_addshmem_write         0  Y  0                          
create_fastlinks            0  -  0                          
dbc_max_pct                 7  -  7                          
dbc_min_pct                 5  -  5                          
default_disk_ir             0  -  0                          
desfree                     -  -  0                          
disksort_seconds            0  -  0                          
dmp_rootdev_is_vol          0  -  0                          
dmp_swapdev_is_vol          0  -  0                          
dnlc_hash_locks           512  -  512                        
dontdump                    0  -  0                          
dskless_node                -  -  0                          
dst                         1  -  1                          
effective_maxpid            -  -  ((NPROC<=30000)?30000:(NPROC*5/4))
eisa_io_estimate            -  -  0x300                      
enable_idds                 0  -  0                          
eqmemsize                  15  -  15                        
executable_stack            0  -  0                          
fcp_large_config            0  -  0                          
file_pad                    -  -  10                        
fs_async                    0  -  0                          
ftable_hash_locks          64  -  64                        
hdlpreg_hash_locks        128  -  128                        
hfs_max_ra_blocks           8  -  8                          
hfs_max_revra_blocks        8  -  8                          
hfs_ra_per_disk            64  -  64                        
hfs_revra_per_disk         64  -  64                        
hp_hfs_mtra_enabled         1  -  1                          
hpux_aes_override           -  -  0                          
initmodmax                 50  -  50                        
io_ports_hash_locks        64  -  64                        
iomemsize                   -  -  40000                      
ksi_alloc_max           64160  -  (NPROC*8)                  
ksi_send_max               32  -  32                        
lotsfree                    -  -  0                          
max_async_ports            50  -  50                        
max_fcp_reqs              512  -  512                        
max_mem_window              0  -  0                          
max_thread_proc          1024  -  1024                      
maxdsiz            0x20000000  -  0X20000000                
maxdsiz_64bit      0x40000000  -  0X40000000                
maxfiles                 2048  -  2048                      
maxfiles_lim             2048  Y  2048                      
maxqueuetime                -  -  0                          
maxssiz             0xa000000  -  0XA000000                  
maxssiz_64bit        0x800000  -  0X800000                  
maxswapchunks            8192  -  8192                      
maxtsiz            0x40000000  Y  0X40000000                
maxtsiz_64bit      0x40000000  Y  0X40000000                
maxuprc                  7218  Y  7218                      
maxusers                 1000  -  1000                      
maxvgs                     10  -  10                        
mesg                        1  -  1                          
minfree                     -  -  0                          
modstrmax                 500  -  500                        
msgmap                   6642  -  (2+MSGTQL)                
msgmax                   8192  Y  8192                      
msgmnb                  16384  Y  16384                      
msgmni                   2048  -  2048                      
msgseg                   6640  -  6640                      
msgssz                      8  -  8                          
msgtql                   6640  -  6640                      
nbuf                        0  -  0                          
ncallout                14067  -  (16+NKTHREAD)              
ncdnode                   150  -  150                        
nclist                  16100  -  (100+16*MAXUSERS)          
ncsize                  14308  -  (NINODE+VX_NCSIZE)+(8*DNLC_HASH_LOCKS)
ndilbuffers                30  -  30                        
netisr_priority             -  -  -1                        
netmemmax                   -  -  0                          
nfile                   14849  -  (16*(NPROC+16+MAXUSERS)/10+32+2*(NPTY+NSTRPTY+NSTRTEL))
nflocks                  4096  -  4096                      
nhtbl_scale                 0  -  0                          
ninode                   9188  -  ((NPROC+16+MAXUSERS)+32+(2*NPTY))
nkthread                14051  -  (((NPROC*7)/4)+16)        
nni                         -  -  2                          
no_lvm_disks                0  -  0                          
nproc                    8020  -  (20+8*MAXUSERS)            
npty                       60  -  60                        
nstrpty                    60  -  60                        
nstrtel                    60  -  60                        
nswapdev                   10  -  10                        
nswapfs                    10  -  10                        
nsysmap                 16040  -  ((NPROC)>800?2*(NPROC):800)
nsysmap64               16040  -  ((NPROC)>800?2*(NPROC):800)
o_sync_is_o_dsync           0  -  0                          
page_text_to_local          -  -  0                          
pfdat_hash_locks          128  -  128                        
public_shlibs               1  -  1                          
region_hash_locks         128  -  128                        
remote_nfs_swap             0  -  0                          
rtsched_numpri             32  -  32                        
scroll_lines              100  -  100                        
scsi_max_qdepth             8  Y  8                          
scsi_maxphys          1048576  -  1048576                    
sema                        1  -  1                          
semaem                  16384  -  16384                      
semmap                   2002  -  (SEMMNI+2)                
semmni                   2000  -  2000                      
semmns                   2000  -  2000                      
semmnu                   1000  -  1000                      
semmsl                   2048  Y  2048                      
semume                    500  -  500                        
semvmx                  32767  -  32767                      
sendfile_max                0  -  0                          
shmem                       1  -  1                          
shmmax             0x20000000  Y  0X20000000                
shmmni                    200  -  200                        
shmseg                    120  Y  120                        
st_ats_enabled              0  -  0                          
st_fail_overruns            0  -  0                          
st_large_recs               0  -  0                          
st_san_safe                 0  -  0                          
streampipes                 0  -  0                          
swapmem_on                  1  -  1                          
swchunk                  2048  -  2048                      
sysv_hash_locks           128  -  128                        
tcphashsz                   0  -  0                          
timeslice                  10  -  (100/10)                  
timezone                  420  -  420                        
unlockable_mem              0  -  0                          
vas_hash_locks            128  -  128                        
vnode_cd_hash_locks       128  -  128                        
vnode_hash_locks          128  -  128                        
vol_checkpt_default     10240  -  10240                      
vol_dcm_replay_size    262144  -  (256*1024)                
vol_default_iodelay        50  -  50                        
vol_fmr_logsz               4  -  4                          
vol_max_bchain             32  -  32                        
vol_max_nconfigs           20  -  20                        
vol_max_nlogs              20  -  20                        
vol_max_nmpool_sz     4194304  -  (4*1024*1024)              
vol_max_prm_dgs          1024  -  1024                      
vol_max_rdback_sz     4194304  -  (4*1024*1024)              
vol_max_vol           8388608  -  (8*1024*1024)              
vol_maxio                 256  -  256                        
vol_maxioctl            32768  -  32768                      
vol_maxkiocount          2048  -  2048                      
vol_maxparallelio         256  -  256                        
vol_maxspecialio          256  -  256                        
vol_maxstablebufsize      256  -  256                        
vol_min_lowmem_sz      524288  -  (512*1024)                
vol_mvr_maxround          256  -  256                        
vol_nm_hb_timeout          10  -  (10)                      
vol_subdisk_num          4096  -  4096                      
vol_vvr_transport           1  -  (1)                        
vol_vvr_use_nat             0  -  (0)                        
volcvm_cluster_size        16  -  16                        
volcvm_smartsync            1  -  1                          
voldrl_max_drtregs       2048  -  2048                      
voldrl_min_regionsz       512  -  512                        
voliomem_chunk_size     65536  -  (64*1024)                  
voliomem_maxpool_sz   4194304  -  (4*1024*1024)              
voliot_errbuf_dflt      16384  -  16384                      
voliot_iobuf_default     8192  -  8192                      
voliot_iobuf_limit     131072  -  131072                    
voliot_iobuf_max        65536  -  65536                      
voliot_max_open            32  -  32                        
volraid_rsrtransmax         1  -  1                          
vps_ceiling                16  -  16                        
vps_chatr_ceiling     1048576  -  1048576                    
vps_pagesize                4  -  4                          
vx_fancyra_enable           0  -  0                          
vx_maxlink              32767  -  32767                      
vx_ncsize                1024  -  1024                      
vx_ninode                   0  -  0                          
vxfs_max_ra_kbytes       1024  -  1024                      
vxfs_ra_per_disk         1024  -  1024                      
vxtask_max_monitors        32  -  32                      


Someone gave me a list of settings that  Oracle recommended for
kernel parameteras follows :
Parameter       Recommended Formula or Value
ksi_alloc_max   (nproc*8)
executable_stack        0
max_thread_proc         1024
maxdsiz         1073741824 (1 GB)
maxdsiz_64bit   2147483648 (2 GB)
maxssiz         134217728 (128 MB)
maxssiz_64bit   1073741824 (1 GB)
maxswapchunks   16384
maxuprc         ((nproc*9)/10)
msgmap          (msgtql+2)
msgmni          (nproc)
msgseg          32767
msgtql          (nproc)
ncsize          (ninode+1024)
nfile           (15*nproc+2048)
nflocks         (nproc)
ninode          (8*nproc+2048)
nkthread        (((nproc*7)/4)+16)
nproc           4096
semmap          (semmni+2)
semmni          (nproc)
semmns          (semmni*2)
semmnu          (nproc-4)
semvmx          32767
shmmax          The size of physical memory or 1073741824 (0X40000000),
               whichever is greater.
               Note: To avoid performance degradation, the value should
               be greater than or equal to the size of the available
               memory.
shmmni          512
shmseg          120
vps_ceiling     64

0
Comment
Question by:sunhux
  • 4
  • 3
  • 2
9 Comments
 
LVL 40

Assisted Solution

by:mrjoltcola
mrjoltcola earned 250 total points
ID: 24781464
I have assumed this is an OLTP system (web system or other).

For high volume systems, this is an observable trend, as the number of parallel sessions increases, the actual response time of queries in those sessions starts to degrade, and so the session windows increase (the time in which it takes to service a session). You can then see an exponential curve of number of processes / sessions.

Its sort of like traffic patterns in real life, on the freeway. Things may flow fine at 70mph until there is some sort of accident, but even after the accident is cleared, there may be a jam for an hour. When things slow down, the avg speed of traffic doesn't just "slightly" decrease, it may fall to 5mph. The same applies to high volume transaction systems.

During those times of peak demand, is 1000 processes enough? Maybe not. You may not be able to realistically tune the single server to handle those spikes, no matter how many processes you allow. Simplye due to the fact that often response time does not necessarily degrade in a linear curve, it often takes an exponential curve.

What can you do about it?

The basic concept in OLTP is move them through the queue as fast as possible.

1) I try to determine how to increase the speed at which the sessions are serviced in the first place, so I decrease parallel session overlap. If you can find highly utilized queries that are taking 0.5 seconds to execute and tune them to 0.1 seconds, by using various Oracle tuning tools and structures, you can possibly scale your system 2, 3, 5, 10x the amount of sessions that can be serviced.

2) With only 4GB of RAM, 1200 process is a lot for Oracle. I would try doubling that RAM and tuning the SGA to cache more things to increase response time. With  HPUX you should be able to get a 3-4 GB SGA easy.

3)  You may consider using Oracle's multi-threaded server (MTS) to reduce the overall number of processes on the system. Due to the high number of processes, your system is a good candidates for MTS. If you do consider this, you should always test the configuration in a test environment.

4) If all else fails, it might be time to consider RAC, so you could simply add a 2nd box of same or better horsepower and scale horizontally without losing your current investment.

5) And then the old solution, upgrade your hardware to faster CPU, more memory

6) You should definitely use the AWR / ADDM reports in Enterprise Manager to observe the system trends surrounding the times when it spikes.


My gut feeling is you are simply running without enough horsepower in reserve for the given transaction response times your system has, and when the load spikes come, your box cannot handle it. Either decrease the response time or increase the horsepower.

0
 

Author Comment

by:sunhux
ID: 24782644
Hi

Actually the users access a Windows Webserver and this webserver retrieves its
data from our HP-UX Oracle Db server.

When I ran top at the times when number of Oracle sessions surge from less than 100
to hit our current limit of 900, the CPU and free memory available is still rather comfortable
(easily 1.5Gb of RAM free with CPU of about 35%)

So we probably won't need hardware upgrade.  Do you think this is an application issue
or some tuning needed?
0
 
LVL 40

Accepted Solution

by:
mrjoltcola earned 250 total points
ID: 24782708
You cannot necessarily use CPU and memory as the indicator for a database server. Often the IO gets maxed before those. At that same time what are your IO stats looking like. If there is an IO bottleneck then adding more memory to the SGA, plus tuning the SQL, more indexes, etc. will still help.

0
 
LVL 47

Assisted Solution

by:schwertner
schwertner earned 250 total points
ID: 24784405
Put in SQLNET.ORA file the parameter

SQLNET.EXPIRE_TIME = 20

20 stands for 20 minutes - you can use another interval.

This will clean the dead sessions.

After that share your observations.
0
Do You Know the 4 Main Threat Actor Types?

Do you know the main threat actor types? Most attackers fall into one of four categories, each with their own favored tactics, techniques, and procedures.

 

Author Comment

by:sunhux
ID: 24793220
Thanks MrJolt.


Hi Mr Schwertner,

All these years, sqlnet.ora never existed, so on 15 June 09, I created sqlnet.ora
which has the following 3 entries :
SQLNET.EXPIRE_TIME = 10
NAMES.DEFAULT_DOMAIN = WORLD
NAME.DEFAULT_ZONE = WORLD

Unfortunately on 30th June, the number of Oracle sessions shot up from below 100
to 900 suddenly on that 30th June's morning and we restarted Oracle.  After that,
the Filenet Storage Manager was not able to connect to Oracle till we removed
the last 2 lines above (ie ... = WORLD).  I was ordered to remove away
sqlnet.ora as part of the Change Control procedure.  My fault that I copied over
sqlnet.ora from another HP-UX which did not face this problem but forgot to remove
the 2 last lines (=WORLD).

We're currently monitoring "sar -v 1 1"  to ensure ninode, nfile & nproc doesn't get
exceeded & on that fateful 30th June morning, the values never get close to the
limits set.

We're running Oracle 9i (reported by tnsping as Version 9.2.0.8.0) but I'm not sure if
the following post from another forum is applicable to our case :

I am not sure of the version of Oracle that you are using but there was a problem I found with the NLB language files.  There is a perl script that is supposed to fix a bug where the broken ones are replaced but it does not run properly.  The one I found was LX400300.NLB and it had a problem internally that was causing Oracle to open it hundreds of times.  If you find the process with lots of files open, you can run lsof (http://hpux.connect.org.uk/hppd/hpux/Sysadmin/lsof-4.82/  if you do not have it on HP-UX media) against the processes.  If you have more than a few, we had 65k of them when nfile was set to 128k, then you know you may have this bug.
 
Oracle has this fix 5257698 with MetaLink article 292492.1. It was a problem with 9i NLB files and how some processes still want to use them instead of the 10g ones.  The perl script were supposed to do these steps below.  Here is one Oracle MetaLink article, registration required, ORA-27041 SVR4 Error: 24: Too many open files after 10.2.0.3 upgrade and patch 5257698 that covers this pretty well with links to other pages.  Note that it was not fixed in 10.2.0.4 or later versions not because the files were not fixed but because the perl script to copy the files was broken and did nothing.
 
Here is what I found in the article:
 
The exact steps to create the "9i behavior" file set:
-----------------------------------------------------
 1) Automatically (If you have perl installed)
  (shutdown all oracle processes first)
  run the $ORACLE_HOME/nls/data/old/cr9idata.pl script to
   create the $ORACLE_HOME/nls/data/9idata directory.

2) Manually (shutdown all oracle processes first) a) make a new subdirectory under $ORACLE_HOME/nls/data called "9idata" b) COPY all files in  $ORACLE_HOME/nls/data/ to  $ORACLE_HOME/nls/data/9idata c) COPY all files from  $ORACLE_HOME/nls/data/old to $ORACLE_HOME/nls/data/9idata   you will need to overwrite some files

(The $ORACLE_HOME/nls/data/old directory only contains the *changed* files, so you need to overwrite a subset of the 500 or more files in $ORACLE_HOME/nls/data/9idata  with the ones found in $ORACLE_HOME/nls/data/old  )
The best way tot check for this being the Oracle problem, other than the lsof above, is to look at $ORACLE_HOME/nls/data/9idata and $ORACLE_HOME/nls/data/old and see what the dates on the files are.  When I looked I found that the files in the $ORACLE_HOME/nls/data/9idata were ancient and from 2006 where the ones in the old directory were from 2008.  The ones in $ORACLE_HOME/nls/data were also from 2008.  Once I was able to provide the proper documentation, Oracle MetaLink articles and lsof output, the DBAs made the change on a test environment before confirming that the problem was resolved.
 
Not sure this is your problem but it sounds just like what I found.  The process keeps opening the NLB file over and over again until it hits the OS limit.  Hope this helps.
 
   -Jim
0
 

Author Comment

by:sunhux
ID: 24793254

Though with "SQLNET.EXPIRE_TIME = 10" in sqlnet.ora, we still get the number
of Oracle sessions surging (it was like that before that too), I'm thinking whether
it's because after implementing this in sqlnet.ora on 15th June, did I shutdown
Oracle completely and restarted the Oracle instances?  Or do I need to restart
the listener?

As for the suggestion from Jim from another forum, it does not appear to be so
because ever since we increased nfile, nproc, the values never come close
to the limits of these HPUX parameters on those days the Oracle processes
hit the limit, so probably Jim's suggestion may not be the right one
0
 
LVL 47

Assisted Solution

by:schwertner
schwertner earned 250 total points
ID: 24797641
Restart the Listener:

lsnrctl stop
lsnrctl start

That parameter will drop only dead sessions.
If the application creates living sessions then it will
keep them. So you have to investigate who creates them.
Some bad parametrized Appl. Server create thousands connections
and use only ten. Investigate the middletier.



Are you that guy that asked us in SQLNET.ORA could be deleted?
0
 

Author Comment

by:sunhux
ID: 24802511
Hi Mr Schwertner,

I did post something on 14th Jun to EE & you replied but I did not ask if
sqlnet.ora could be deleted;  rather all these years it was not absent in
this particular problem HP-UX server (before I joined the company).
I've appended furthest below the original posting on 14th Jun.

Just found there's a daily cronjob that restarts oracle+listener at night
so "SQLNET.EXPIRE_TIME = 10"  on 14-15th Jun would have taken effect
from 15-30 June but we still encountered Oracle processes surging to
reach the limit of 900 on 30th June.

Following is extracted from listener daily log which showed it's being
restarted, in case it gave any clue :

Trace information written to /oracle/Ora92/network/trace/ora9ilstnr.trc
Trace level is currently 0
Started with pid=9829  <==
Listening on: (DESCRIPTION=(ADDRESS=(PROTOCOL=tcp)(HOST=XXXHP16)(PORT=1521)))
. . . . .
(HOST=XXXHP16)(USER=filenet))) * (ADDRESS=(PROTOCOL=tcp)(HOST=10.51.1.56)(PORT=52793)) * establish * idb * 0
08-JUL-2009 06:42:41 * (CONNECT_DATA=(SERVICE_NAME=idb)(CID=(PROGRAM=shf)(HOST=SLAHP16)(USER=filenet))) * (ADDRESS=(PROTOCOL=tcp)(HOST=10.51.1.56)(PORT=52912)) * establish * idb * 0
 . . . (lots of User=filenet connection like above line) . . .


Following parameters comparing a non-problem HP-UX with the probelm HP-UX :
On our "good" Itanium HPUX running Oracle+SAP :
semmni                   2000  -  2000  (suggest to be = nproc; currently nproc is 8020)
semmns                   2000  -  2000 (=semmni x2 ?)
semmnu                   1000  -  1000 (suggest to be nproc-4)
vps_ceiling                      64  64

On our "problem" PARISC HPUX running Oracle+Filenet+App
semmni                         8400  8400
semmns                        16800  16800
semmnu                         4196  4196
vps_ceiling                          16  -  16


========================================================

Finally, the original posting to EE on 14th June follows :




Hi

For the last couple of months, our HP-UX  B11.11 PA-RISC kernel parameter
nfile (possibly nproc) limits were hit (in syslog, the message was
"May 21 10:25:00 hostname vmunix: file: table is full"),
so we increased maxusers which in turn automatically increases nfile, nproc etc.


Things stabilized for about 3-4 weeks & the users (running IE browser to
'browse'  a Wintel IIS server which in turn retrieves its data from our
Unix DB server.  This Unix DB server has tnsnames.ora but not sqlnet.ora:
I don't know why as I was told by my predecessor, there's been no
requirement to have sqlnet.ora (besides this server has been around for
four years and only starting last year this problem surfaced).

Error logs from lsnrdb.log :

=========================================


lsnridb-20090615-1101.log:  TNS-12560: TNS:protocol adapter error
lsnridb-20090615-1101.log:    HPUX Error: 11: Resource temporarily unavailable
lsnridb-20090615-1101.log:  TNS-12560: TNS:protocol adapter error
lsnridb-20090615-1101.log:    HPUX Error: 11: Resource temporarily unavailable
lsnridb-20090615-1101.log:  TNS-12560: TNS:protocol adapter error
lsnridb.log:Error listening on: (DESCRIPTION=(ADDRESS=(PROTOCOL=TCP)(HOST=MLAHP16)(PORT=1521)))
lsnridb.log: TNS-12560: TNS:protocol adapter error
lsnridb.log:   HPUX Error: 226: Address already in use

===================

15-JUN-2009 10:35:25 * (CONNECT_DATA=(SID=idb)(CID=(PROGRAM=c:\windows\system32\inetsrv\w3wp.exe)(
HOST=NTSLA56)(USER=JDSAppAdmin))) * (ADDRESS=(PROTOCOL=tcp)(HOST=10.51.21.116)(PORT=3442)) * establ
ish * idb * 0
15-JUN-2009 10:35:25 * (CONNECT_DATA=(SID=idb)(CID=(PROGRAM=c:\windows\system32\inetsrv\w3wp.exe)(
HOST=NTSLA56)(USER=JDSAppAdmin))) * (ADDRESS=(PROTOCOL=tcp)(HOST=10.51.21.116)(PORT=3443)) * establ
ish * idb * 12500
TNS-12500: TNS:listener failed to start a dedicated server process
 TNS-12540: TNS:internal limit restriction exceeded
  TNS-12560: TNS:protocol adapter error
   TNS-00510: Internal limit restriction exceeded
    HPUX Error: 11: Resource temporarily unavailable
15-JUN-2009 10:35:25 * (CONNECT_DATA=(SID=idb)(CID=(PROGRAM=c:\windows\system32\inetsrv\w3wp.exe)(
HOST=NTSLA56)(USER=JDSAppAdmin))) * (ADDRESS=(PROTOCOL=tcp)(HOST=10.51.21.116)(PORT=3444)) * establ
ish * idb * 12500
TNS-12500: TNS:listener failed to start a dedicated server process
 TNS-12540: TNS:internal limit restriction exceeded
  TNS-12560: TNS:protocol adapter error
   TNS-00510: Internal limit restriction exceeded
    HPUX Error: 11: Resource temporarily unavailable
15-JUN-2009 10:35:25 * (CONNECT_DATA=(SID=idb)(CID=(PROGRAM=c:\windows\system32\inetsrv\w3wp.exe)(
HOST=NTSLA56)(USER=JDSAppAdmin))) * (ADDRESS=(PROTOCOL=tcp)(HOST=10.51.21.116)(PORT=3445)) * establ
ish * idb * 12500
TNS-12500: TNS:listener failed to start a dedicated server process


This is Oracle 9i

Any idea to go about solving this and let me know if you need more logs


Then users encountered the same error again last couple of days but for
this time round, users' sqlnet logs have the same error but on Unix Db
server end, the syslog were uneventful (nothing useful & no previous
Their Oracle sqlnet client error message is :

9697-2883



user Boey :
[3/30/2009 10:11:19 AM] [JDS.DBACTIONS.JDSORACLE] [Fill(qryText;
aUserObject)] [Boey Ken Lok] [ERR] [Select EPROCESS_ROUTER,
EPROCESS_HOST, EPROCESS_PORT FROM JDS_S_SYSTEMPARAM (Inner) ORA-12500:
TNS:listener failed to start a dedicated server process]
[3/30/2009 10:22:14 AM] [JDS.DBACTIONS.JDSORACLE] [Fill(qryText;
aUserObject)] [Boey Ken Lok] ] [ERR] [Select EPROCESS_ROUTER,
EPROCESS_HOST, EPROCESS_PORT FROM JDS_S_SYSTEMPARAM (Inner) ORA-12571:
TNS:packet writer failure]



sunhux:

Hi again,
is it
  SQLNET.TIME_EXPIRE
or
  SQLNET.EXPIRE_TIME
0
 
LVL 47

Assisted Solution

by:schwertner
schwertner earned 250 total points
ID: 24803807
SQLNET.EXPIRE_TIME


sqlnet.expire_time is useless against JDBC connections.
why not use MTS with connection pooling , but MTS is not good for
connections which hold session connection for a long time.
0

Featured Post

How to run any project with ease

Manage projects of all sizes how you want. Great for personal to-do lists, project milestones, team priorities and launch plans.
- Combine task lists, docs, spreadsheets, and chat in one
- View and edit from mobile/offline
- Cut down on emails

Join & Write a Comment

This post first appeared at Oracleinaction  (http://oracleinaction.com/undo-and-redo-in-oracle/)by Anju Garg (Myself). I  will demonstrate that undo for DML’s is stored both in undo tablespace and online redo logs. Then, we will analyze the reaso…
Every server (virtual or physical) needs a console: and the console can be provided through hardware directly connected, software for remote connections, local connections, through a KVM, etc. This document explains the different types of consol…
This video shows information on the Oracle Data Dictionary, starting with the Oracle documentation, explaining the different types of Data Dictionary views available by group and permissions as well as giving examples on how to retrieve data from th…
This video shows how to copy an entire tablespace from one database to another database using Transportable Tablespace functionality.

743 members asked questions and received personalized solutions in the past 7 days.

Join the community of 500,000 technology professionals and ask your questions.

Join & Ask a Question

Need Help in Real-Time?

Connect with top rated Experts

13 Experts available now in Live!

Get 1:1 Help Now