Swap utilization %
Hi,
I am running 2 database instances of Oracle 11g on Solaris 10 (SPARC T-5120).
At the OS level it seems that the swap is not being used at all, but at the OEM, it shows the "Swap Utlization %" is about 80%
bash-3.00$ swap -l
swapfile dev swaplo blocks free
/dev/dsk/c1t0d0s3 32,3 16 36877808 36877808
Could someone please help me to understand these two indicator and to which one I need to pay attention.
Also, I noticed that if I increase the "memory_max_target" parameter, the "Swap Utilization%" increases.
Regards,
AShum
ashum wrote:
I don't know how to format it to make it more readable on the post.
Any advice?
Thanks,
Alberto
SQL and PL/SQL FAQ
scroll down to #9 & do as it says
Find my second post in this thread which contains the line below
What columns are produced by your version of vmstat?What I posted was from
man vmstat
that documents the output from the vmstat command.
you need to do the same
man vmstat
then COPY the actual description for each column of the output from vmstat & PASTE all here
be sure to wrap the text in tags
Similar Messages
-
Swap utilization is 100 %
Hi All,
In my production system swap utilization is 100% that is causing performance issues, when we execute top & Glance command , couldn't find workprocess which are using more memoery here i could see only 2 process are using 2 gb memory, actually we have 32GB memoery. also we have checked in sm04, ther also no workprocess is using much memory. In St02 detailed analysis also utilizing memory is 3gb. i could see occupied physical memoery in ST06. there only 1GB free meomery.
Could you please some one help me how we can find where this physicall memory is utilizing.What's your current physical memory size?
Just set all memory related parameters according to your physical memory and re-check.
Increase your physical memory if required (contact SAP for fine tuning)
Regards,
Nick Loy -
EM alert Message : Swap Utilization is 100%
Hi,
on 10G R2, EM sends us the following alert :
Message=Swap Utilization is 100%And in details when I look, it says : examine the applications that do not belong to this instance.
I wonder if SGA and PGA are well tunned ?
I hyave the followings :
SQL> show sga
Total System Global Area 1476395008 bytes
Fixed Size 1251172 bytes
Variable Size 293603484 bytes
Database Buffers 1174405120 bytes
Redo Buffers 7135232 bytes
SQL> show parameter target
NAME TYPE VALUE
archive_lag_target integer 0
db_flashback_retention_target integer 2880
fast_start_io_target integer 0
fast_start_mttr_target integer 0
pga_aggregate_target big integer 1100M
sga_target big integer 5504MMany thanks.Hi,
today RMAN backup failed with :
ORA-04030 :out of process memory when trying to allocate 2457618 bytes (pga heap,zbits_kgcstate)might this error be related to EM alert :
Swap Utilization IS 100%
according to mrtalink note 373602.1 we should be in automatic SGA to avoid ORA-04030 , and we are then why ORA-04030 ?
SQL> SHOW parameter target
NAME TYPE VALUE
archive_lag_target integer 0
db_flashback_retention_target integer 2880
fast_start_io_target integer 0
fast_start_mttr_target integer 0
pga_aggregate_target big integer 1100M
sga_target big integer 5504MThanks before. -
Swap Utilization is 100% - cant clear alert
We have had the Swap Utilization is 100%, crossed warning (35) or critical ( ) threshold hanging around for a month now and I cant clear it.
The server has been rebooted and there is no swapping.
I found this http://www.ora-solutions.net/web/2010/11/12/grid-control-11g-agent-metric-swap-utilization-on-hp-ux-with-pseudo-swap/
but its not for our platform nor our Grid Agent version!
Any ideas please?
Linux RHEL 5
Oracle Enterprise Manager 10g Release 5 Grid Control 10.2.0.5.0If you find that Metric via the host that it is configured on, just clear the fields that are there for that metric.
Then when you can see those alerts cleared, re-enter the values and carry on regardless.
DA -
Oracle Coherence increasing Swap Utilization
We are using Oracle Coherence on linux servers. However, we noticed that because of Coherence processes running, often our swap utilization % increases too much, sometimes becoming more than 98%, even touched 100% a few times.
Once we kill all the coherence related processes, then it becomes normal.
Is there any way we can make coherence processes to only use a particular size of Swap space ?
Currently increasing swap space is not in our scope.
Please suggest.
Edited by: user7761515 on May 3, 2012 11:29 AMHi,
We are using Oracle Coherence on linux servers. However, we noticed that because of Coherence processes running, often our swap utilization % increases too much, sometimes becoming more than 98%, even touched 100% a few times.
Swapping itself (1%-100%) is not a good sign and should be avoided by ensure that you have sufficient memory such that you are not making active use of swap space on your machines. The active usage of SWAP space will have significant performance degradation.
Is there any way we can make coherence processes to only use a particular size of Swap space ?Manage your memory by allocating heap using -Xmx for Coherence JVMs. You need to ensure that the sufficient RAM memory is available on the server for Coherence JVMs and other operating system processes and do not consume all the RAM.
To temporarily set the swappiness, as the root user echo a value onto /proc/sys/vm/swappiness. The following command will set swappniess to 0:
echo 0 >/proc/sys/vm/swappiness //To set the value permanently, modify the /etc/sysctl.conf file.
Hope this helps!
Cheers,
NJ -
Oracle Linux - Swap Utilization
I am running Oracle 11.2.0.2 on Oracle Linux x86 64-bit.
In looking at OEM, I see my swap utilization reported at 17.01% and I see virtual memory paging listed in the ADDM Performance and Analysis section:
Host operating system was experiencing significant paging but no particular root cause could be detected. Investigate processes that do not belong to this instance running on the host that are consuming significant amount of virtual memory. Also consider adding more physical memory to the host.
There no additional or non-standard processes running on this server.
For those of you that are running similar configurations, does this seem normal or typical for you? Might you guess that I have potential problems with swapping or is this normal or an EOM “false alarm”?
Additional info:
My server is virtual. I am using Huge Pages and I am certain my SGA is fully contained in there.
Memory Statistics
Host Mem (MB): 14,031
SGA use (MB): 8,192
PGA use (MB): 509
% Host Mem used for SGA+PGA: 62.01
Thanks for your time.below takes 1 minute to complete
as long as (so+si) less than (bo+bi) then RAM is NOT a bottleneck
bcm@bcm-laptop:~$ vmstat 6 10
procs -----------memory---------- ---swap-- -----io---- -system-- ----cpu----
r b swpd free buff cache si so bi bo in cs us sy id wa
0 0 9776 105876 290164 1935116 0 0 33 38 228 300 9 4 86 1
2 0 9776 103364 290172 1937504 0 0 19 423 700 1585 23 3 70 4
0 0 9776 118616 290176 1940692 0 0 0 28 962 2483 10 5 85 0
0 0 9776 113796 290180 1940728 0 0 3 126 941 2355 10 4 85 1
1 0 9776 113796 290220 1940696 0 0 0 86 786 1779 4 3 91 3
0 0 9776 113704 290236 1940696 0 0 0 20 812 1839 4 3 92 1
1 0 9776 113704 290300 1940708 0 0 1 81 755 1706 4 2 92 2
0 0 9776 108604 290324 1940716 0 0 1 56 770 1810 5 2 92 1
0 0 9776 108108 290340 1940752 0 0 1 1681 790 1797 3 3 89 5
0 0 9776 108108 290356 1940752 0 0 0 120 851 1736 4 2 92 1
bcm@bcm-laptop:~$ -
Swap utilization problem?
Hi
i have use Oracle 10g Database control, in everyday i screen shot the performance image (24 hours)
but yesterday i have screen shot image ,the memory untilization is 100%
how can utilizaion is 100% ? and what is solutionHi,
'how can utilizaion is 100% ? and what is solution '
Well this is a question on which you should give an answer.
You have to know what was going on in your database.
Using Dbcontrol i would use AWR feature and work with the snapshot saved in that period.
Acr -
Swap utilization metric under HP/UX
Hello,
We are using OEM Grid to manage to manage some 10g databases running on HP-UX B.11.11 U (64).
WNo idea???
-
What is normal swap space utilization on Solaris 10
Hi all,
I'm running Oracle 11.2 on Solaris 10 on a couple of HP Proliant DL 360 servers.
Both servers have 72G of physical RAM with swap space set to 16G on both of them.
Server A has only one database and total memory free = 30G.
Server A
top: Memory: 72G phys mem, 30G free mem, 16G total swap, 16G free swap
swap -s: total: 27249744k bytes allocated + 13873764k reserved = 41123508k used, 1000552k available
prstat:
NPROC USERNAME SWAP RSS MEMORY TIME CPU
257 oracle 39G 38G 53% 222:11:52 5.6%
31 root 57M 59M 0.1% 414:47:23 0.1%
1 smmsp 1776K 7736K 0.0% 0:00:34 0.0%
6 zabbix 4752K 4092K 0.0% 0:58:31 0.0%
4 daemon 3864K 6456K 0.0% 0:00:35 0.0%Server B has two databases and total memory free = 9G.
Server B
top: Memory: 72G phys mem, 9890M free mem, 16G total swap, 16G free swap
swap -s: total: 29223360k bytes allocated + 627312k reserved = 29850672k used, 16926320k available
prstat:
NPROC USERNAME SWAP RSS MEMORY TIME CPU
157 oracle 28G 28G 39% 15:38:41 0.4%
34 root 58M 65M 0.1% 2:56:57 0.0%
6 zabbix 5580K 4816K 0.0% 0:00:31 0.0%
1 smmsp 1776K 5724K 0.0% 0:00:00 0.0%
5 hpsmh 17M 13M 0.0% 0:00:00 0.0%
4 daemon 3204K 5912K 0.0% 0:00:00 0.0%We are using zfs file system on both servers (which is pretty much the standard these days on Solaris).
Recently I got an OEM alert that my swap space on server A had crossed the 95% threshhold on one of the servers.
But when I checked the server, I found that the average swap space utilization was 97.45.
In fact, what actually happened was my swap utilization momentarily dropped below 95% and then returned back to its normal range above 95% which caused the alert to be triggered.
So this made me wonder why my swap space utilization was so high on server A, or is this just normal for Solaris (v.10).
Checking with server B, I see that my swap utilization is only at 63.6% (even though server B has much more physical memory in use by the two databases than server A).
Main question is why is swap utilization so high on server A, which is configured the same as server B and with less physical memory actually in use.
Next question is should I be concerned.
When I check vmstat, I do not see any paging in or out or blocked processes.
See below for server A
Server A
$ vmstat -S 5 5
kthr memory page disk faults cpu
r b w swap free si so pi po fr de sr s0 s1 s2 s5 in sy cs us sy id
0 0 0 1059868 30507176 0 0 0 0 0 0 2 7 -0 123 30 13742 25008 7264 5 2 93
0 0 0 1024076 30982140 0 0 0 0 0 0 0 23 0 0 122 4433 14793 6854 6 2 92
0 0 0 1030292 30987500 0 0 0 0 0 0 0 0 0 0 102 4055 15049 7014 8 1 91
0 0 0 1044484 30999572 0 0 0 0 0 0 0 0 0 0 129 5905 19196 8127 6 1 93
0 0 0 1028584 30987636 0 0 0 0 0 0 0 0 0 0 114 10611 19925 7259 6 3 90974632 wrote:
Looks like we don't have 'free' on these Solaris boxes (only the man pages).
I'm guessing that free is for linux (since it works fine on my linux boxes).
$ whereis free
free: /usr/man/man3c/free.3cdarn!
Realize that SWAP is purely an OS facility; which is 100% external to Oracle.
The OS send little used or idle processes into SWAP when RAM is scarce resource.
The fact that SWAP is being used is not a Bad Thing, in and of itself.
as long as vmstat shows that BI+BO > SI+SO I would ignore the Chicken Little warnings. -
Excess Swap Space....
Hi all,
Please I would like to find out if large swap space has any effect on the performance of a Server.
ThanksNo. Generally not a good idea.
Oh, details? Sure.
To begin with you do <u>not</u> want any active paging going on; it kills performance quickly. It is normal to see some memory pages being evicted to paging store; the kernel does some proactive evictions of stale memory pages just in case the room is needed sometime in the future.
$ top -n1
top - 07:37:55 up 3 days, 19:58, 1 user, load average: 0.01, 0.82, 1.26
Tasks: 302 total, 15 running, 287 sleeping, 0 stopped, 0 zombie
Cpu(s): 4.5%us, 6.2%sy, 0.1%ni, 88.2%id, 0.8%wa, 0.0%hi, 0.1%si, 0.0%st
Mem: 6902244k total, 6486860k used, 415384k free, 600968k buffers
Swap: 8392696k total, 14496k used, 8378200k free, 1290552k cachedThe swap utilization reported by top(1) is misleading because it shows a static picture of how much swap space is being used. The problem is that you can't distinguish between the stale-memory evictions and the unwanted paging activity. Better to use the vmstat(1) tool for this:
$ vmstat 1 3
procs -----------memory---------- ---swap-- -----io---- --system-- -----cpu-----
r b swpd free buff cache si so bi bo in cs us sy id wa st
1 0 14496 411440 601624 1293196 0 0 48 103 28 24 5 6 88 1 0
61 0 14496 411316 601632 1293196 0 0 0 212 5224 6751 2 4 93 2 0
5 0 14496 411152 601640 1293188 0 0 0 16 7205 8826 5 11 83 1 0The interesting metric here is the swap-in (si) column, which counts the number of pages read into memory from the paging store. It is OK for this to show some minor values, but should the count increase rapidly for some time you should regard this as a sign of swap activity and fix it.
Lots of installation documentation suggests that the size of paging store be at least equal to the physical memory size or even larger. On anything much larger than an old desktop this is simply too much space: the system will become unusably slow long before this amount of space is utilized.
Ancient history: at one time some UNIX ports used the swap space to temporarily store a memory image of a system oops. Every boot, the kernel checked the swap space for a special signature pattern to see if the swap space held such an image.
Back to the present: having an excessively-large paging store simply wastes disk space. Sometimes the extra paging store can be useful to have because it can postpone a Linux out-of-memory (OOM) kill by the kernel if you have a run-away program leaking memory, but most likely this simply trades an OOM kill for really poor performance. Your mileage may vary, as they say. -
Reg LGWR got killed by OS utility
Hi,
Last saturday i faced a strange issue in my database.
My environment:
Oracle EBS.
Two node installation.
Forms and Web in One server and
Concurrent and Database in Another.
We have sheduled backup on every saturday.
Duing last saturday,after backup was completed,We started the instance it was normal,
We also checked the working on instance and concurrent managers by running reports.
it was around 9 pm.
Then sysadmin guy has started restoration from san storage using tivoli client to one mount point in
same database server.
It was around 10.24 pm LGWR process was killed automatically by OOM(out of memory killer) in Linux.
I have copied the alert log entries during that time.
DETAILS FROM bdump from Oracle Database
Sat Jun 5 20:58:20 2010
Starting ORACLE instance (normal)
PMON started with pid=2, OS id=13375
PSP0 started with pid=3, OS id=13377
MMAN started with pid=4, OS id=13379
DBW0 started with pid=5, OS id=13381
DBW1 started with pid=6, OS id=13383
LGWR started with pid=7, OS id=13385
CKPT started with pid=8, OS id=13387
SMON started with pid=9, OS id=13389
RECO started with pid=10, OS id=13391
CJQ0 started with pid=11, OS id=13393
MMON started with pid=12, OS id=13395
MMNL started with pid=13, OS id=13397
Sat Jun 5 22:24:42 2010
Errors in file /oracle/PROD/app/proddb/10.2.0/admin/PROD_dbserv1/bdump/prod_pmon_13375.trc:
ORA-00470: LGWR process terminated with error
Sat Jun 5 22:24:45 2010
Errors in file /oracle/PROD/app/proddb/10.2.0/admin/PROD_dbserv1/bdump/prod_ckpt_13387.trc:
ORA-00470: LGWR process terminated with error
Sat Jun 5 22:24:48 2010
PMON: terminating instance due to error 470
Sat Jun 5 22:24:48 2010
Errors in file /oracle/PROD/app/proddb/10.2.0/admin/PROD_dbserv1/bdump/prod_dbw0_13381.trc:
ORA-00470: LGWR process terminated with error
Sat Jun 5 22:24:48 2010
Errors in file /oracle/PROD/app/proddb/10.2.0/admin/PROD_dbserv1/bdump/prod_psp0_13377.trc:
ORA-00470: LGWR process terminated with error
Sat Jun 5 22:24:48 2010
Errors in file /oracle/PROD/app/proddb/10.2.0/admin/PROD_dbserv1/bdump/prod_mman_13379.trc:
ORA-00470: LGWR process terminated with error
Instance terminated by PMON, pid = 13375
Below are trace file contents.
prod_pmon_13375.trc
*** 2010-06-05 22:24:42.417
*** SERVICE NAME:(SYS$BACKGROUND) 2010-06-05 22:24:38.983
*** SESSION ID:(1000.1) 2010-06-05 22:24:38.983
Background process LGWR found dead
Oracle pid = 7
OS pid (from detached process) = 13385
OS pid (from process state) = 13385
dtp = 0x2000bbf0, proc = 0x52a15e84
error 470 detected in background process
ORA-00470: LGWR process terminated with error
==========
prod_ckpt_13387.trc
*** 2010-06-05 21:34:57.543
*** SERVICE NAME:(SYS$BACKGROUND) 2010-06-05 21:34:56.241
*** SESSION ID:(994.1) 2010-06-05 21:34:56.241
*** 2010-06-05 21:34:56.241 62692 kcrr.c
ARCH: Detected ARCH process failure
*** 2010-06-05 21:40:32.152
*** 2010-06-05 21:40:32.152 62692 kcrr.c
ARCH: Detected ARCH process failure
*** 2010-06-05 21:45:04.723
*** 2010-06-05 21:45:04.723 62692 kcrr.c
ARCH: Detected ARCH process failure
*** 2010-06-05 22:24:45.661
error 470 detected in background process
ORA-00470: LGWR process terminated with error
==================
prod_dbw0_13381.trc
*** 2010-06-05 22:24:48.507
*** SERVICE NAME:(SYS$BACKGROUND) 2010-06-05 22:24:45.800
*** SESSION ID:(997.1) 2010-06-05 22:24:45.800
error 470 detected in background process
ORA-00470: LGWR process terminated with error
===================
prod_psp0_13377.trc
*** 2010-06-05 22:24:48.509
** SERVICE NAME:(SYS$BACKGROUND) 2010-06-05 22:24:46.405
*** SESSION ID:(999.1) 2010-06-05 22:24:46.405
error 470 detected in background process
ORA-00470: LGWR process terminated with error
===================
prod_mman_13379.trc
*** 2010-06-05 22:24:48.510
*** SERVICE NAME:(SYS$BACKGROUND) 2010-06-05 22:24:47.705
*** SESSION ID:(998.1) 2010-06-05 22:24:47.705
error 470 detected in background process
ORA-00470: LGWR process terminated with error
DETAILS FROM /var/log/messages from OS
Jun 5 22:39:24 dbserv2 kernel: Out of Memory: Killed process 27262 (oracle).
Jun 5 22:35:27 dbserv2 kernel: Out of Memory: Killed process 21588 (oracle).
Jun 5 22:30:17 dbserv2 kernel: Out of Memory: Killed process 20745 (oracle).
Jun 5 22:29:59 dbserv2 kernel: Out of Memory: Killed process 20101 (oracle).
Jun 5 22:29:41 dbserv2 kernel: Out of Memory: Killed process 19896 (oracle).
Jun 5 22:29:25 dbserv2 kernel: Out of Memory: Killed process 16243 (oracle).
Jun 5 22:25:23 dbserv2 kernel: Out of Memory: Killed process 13670 (oracle).
Jun 5 22:24:47 dbserv2 kernel: Out of Memory: Killed process 13035 (oracle).
Jun 5 22:24:39 dbserv2 kernel: Out of Memory: Killed process 13385 (oracle). LGWR started with pid=7, OS id=13385
Jun 5 22:24:30 dbserv2 kernel: Out of Memory: Killed process 2436 (oracle).
Jun 5 22:24:30 dbserv2 kernel: Out of Memory: Killed process 2192 (oracle).
Jun 5 22:24:30 dbserv2 kernel: Out of Memory: Killed process 12585 (oracle).
Jun 5 22:23:58 dbserv2 kernel: Out of Memory: Killed process 6985 (oracle).
Jun 5 21:30:47 dbserv2 kernel: Out of Memory: Killed process 20644 (oracle).
If any of you have faced the same issue kindly help me.
Regards
KrishnaFrom OS error message , Looks like there was resource crunch at the time of crash.
What was the CPU/Memory/Swap utilization at time of crash.
Also check out for kernel parameters.
Please follow below MOS note and also raise SR with support. :
Pmon terminated instance due to LGWR termination ORA-470 [ID 431246.1]
BACKGROUND PROCESS DIES, NO TRACE FILE GENERATED [ID 1029808.6]
Regards
Rajesh -
We have oracle 10g running on a HP unix. We want to implement an application deployment before which we want to do performance and load test. How do I find how much memory each session is consuming and how much SGA/PGA are utilized and how much is free as we increase the load. Is there anything else I need to monitor other than memory, CPU utilization, swap utilization etc.?
ThanksYou should do performance and load test against your application. Each application behave differently, the idea is tune your DB to work best with your application not finding an ideal setup before implement your application because there isn't any.
You can find a test box that similar to your production, run your application to mimic the stress load and check AWR report for peak usage time and see if there's anything need improve. -
Memory graphic (OEM 10g)
Hi guys,
One doubt regarding the Oracle enterprise manager and the graphics of memory utilization.
In performance tab I´m seeing memory utilization like 99% of use and swap utilization 3%
if I increase sga_max_size Will I have more memory available in this case ? This grahics shows the memory available to oracle or the full system memory ?
Oracle 10gR1
thanks,
sylari am experiencing this too. memory usage is too high (98%) and i don't know why it is happening. anybody wants to share their idea?
10gR1 on linux
thanks.
Message was edited by:
mervin14 -
Hi Gurus,
I require the mane of table in which i can find the following information
1.)type of system (system info)
2.)client
3.)server model (like NT or UNIX)
4.)Instances
5.)Server Names
6.)Memory info
7.)Operating system
8.)SAP SID's
I only know 2 tables i.e, T000 and CVERS
Can anyone help for the rest.
Regards
Anthony D'souzadear Anthony D'souza,
why using table, if there is tcode provided by SAP to monitor it all ?
here I show you the way
using TCode ST06 > detail analysis menu> HW info
it tells you system model, no. of CPU, disks, tapes, memory, network interface, disk config, swap utilization and kernel parameters
using TCode ST06 > detail analysis menu> Operating System Collector> Details On
it tells you OS type and version, hostname of your server,
SAP SID, client, program currently running and user ID can be seen on the bottom right corner of your GUI screen
hope it help you.
rgds,
alfonsus guritno -
How to monitor /tmp in solaris in OEM 12c
Oracle says by default they don't monitor swap file system and only metric available is swap utilization %
I would like to monitor /tmp and warning alert when it reaches 80% and 95% as a critical alert.
Any thoughts... Oracle suggests to do metric extension and I don't think for a simple thing like this we need to create a metric extension...Check the monitoring under Host for Filesystem Space Available (%), this should be monitoring all filesystems by default.
Maybe you are looking for
-
I am planning a long vacation and taking my laptop but it is my home computer that my new ipad and ipod touch are synced too. I know i can backup to the cloud over wifi but how can i do things like delete photo albums instead of having to manually de
-
BPEL processes not available on BPEL console when deployed from JDeveloper
Installation: SOA Suite 10.1.3.4 (Upgraded through patch) (Standalone OC4J) When I deploy the BPEL processes through JDeveloper, they do not show up in the BPEL console. JDeveloper does not display any error message but in fact displays successful de
-
Where I have to resolve my problem for double billing for an app I purchased. App under question is ' DOCS TO GO' both normal and premium version. Help me Thanks
-
Remove partition key from partitoned table
can somebody help me with a sql syntax to remove a partition key from a partiton table . The scenario is explained as follows: create table t1 (c1 number, c2 number ,c3 number) partition by range (c1, c2) (partition p1 values less than 11, partition
-
I get the JRun Servlet Error when trying to run cf admin. port is 8500 for machine name , i have tried localhost, locathost:localdomain, 127.0.0.1 but none of them work. where am i going wrong? I have installed coldfusion mx 7. In services configurat