[hobbit] dbcheck for RAC connections refused

26 Oct 2009


      Hi Martin,
if you run the dbcheck.pl script from your xymon server you can put the environment directly in the the server/etc/hobbitserver.cfg
if you run it on a client it should be put in the client/etc/hobbitclient.cfg
You can also check the db connectivity before running the script in this way...
With the xymon user define the ORACLE_HOME, ORACLE_SID, TNS_ADMIN environment variable and try to run sqlplus and connect with it.
If it works also dbcheck should work.
To do a real test now you can run dbcheck.pl manually.
The minimum options required in dbcheck.ini are:
[racinstance]
oraclerac=yes
dbtype=oracle
sid=racsid
Not knowing hostname it should use oracle library and tnsnames to reach the database.
Let me know if it works or you have problems...
Francesco
-----Original Message-----
From: Martin Flemming [mailto:martin.flemming at desy.de]
Sent: giovedì 22 ottobre 2009 18.58
To: hobbit at hswn.dk
Subject: RE: [hobbit] dbcheck for RAC connections refused
Dear, Francesco et all !
.. long time ago ...
Please give me an hint how and where i can set the TNS_ADMIN variable ?
i'm a very bad coder :-(
.. various possible (for me) options tested, but without success :-(
thanks in advance
    martin
On Wed, 9 Sep 2009, Francesco Duranti wrote:
...
Hi Martin,
you can try to set the variable for the oracle home into the dbcheck.ini configuration file.
oraclehome              = /usr/local/oracle/10.2.0/client/
It should point to the oracle home under which you have your client
library and network/admin/tnsnames.ora file
I've also checked on the documentation on DBD::Oracle and it should be possible to have the "*.ora" file in a different location. You should set the environment variable TNS_ADMIN to the directory containing tnsnames.ora file but I've not checked if it works. If you're running the dbcheck.pl on the centrally on the xymon server you can define the environment variable in the hobbitserver.cfg or you can also define it in the startup environment variable of your xymon user so that it's defined when the dbcheck.pl script is running.
Let me know if this work for you. I'll also try to do some tests and put a new config parameters on the next version of dbcheck related to the tnsnames.ora location.
Francesco
-----Original Message-----
From: Martin Flemming [mailto:martin.flemming at desy.de]
Sent: mercoledì 9 settembre 2009 11.17
To: hobbit at hswn.dk
Subject: RE: [hobbit] dbcheck for RAC connections refused
Hi, again !
My setup works now for the RAC-instances ... ;-)
of course a configuration-error by me and my oracle-admin ..
But unfortunatley not for the RAC-DB ..
Indeed the message is clear
Wed Sep 09 09:24:46 2009 Connection Check Failed
and
ORA-12154: TNS:could not resolve the connect identifier specified (DBD
ERROR: OCIServerAttach)
Because, i've got an local tnsnames.ora for this RAC-Cluster .. and not our global oracle-enviroment and tnsnames.ora ...
Is there a possibility to make an variable in dbcheck.ini for the  place of the tnsnames.ora ?
thanks & cheers,
   martin
On Wed, 5 Aug 2009, Martin Flemming wrote:
...
Hi !
.. was some days on holiday .. ;-)
Thanks a lot for help !
.. indeed, something is going wrong, with the communication to the
rac-database ..
you see, i'm not a oracle-expert ...
i will contact my second oracle-admin and will ask him, whats going
wrong with the configuration from my first oracle-admin ...
Thanks, i will report ...
cheers,
martin
On Fri, 31 Jul 2009, Francesco Duranti wrote:
...
This is really strange.
Regarding the RAC test it cannot be done without the tnsnames.ora
configuration.
What's the problem regarding the tnsnames.ora? does it exists?
The timeout problem seems also strange because it's like your
client could  not reach the database.
You should try to connect to the oracle client (sqlplus).
You can also try with this perl script
#/usr/bin/perl
use DBI;
$ENV{'ORACLE_HOME'}="/usr/local/oracle/10.2.0/client/";
$dbhost="hostname";
$port=1521;
$sid="SID";
$user="user";
$pw="password";
$data_source="DBI:Oracle:host=$dbhost;port=$port;sid=$sid";
$dbh = DBI->connect($data_source,$user,$pw ,{ RaiseError => 1,
PrintError  =>1});  $dbh->disconnect;  exit 0;
Let me know if there's any problem...
Francesco
-----Original Message-----
From: Martin Flemming [mailto:martin.flemming at desy.de]
Sent: venerdì 31 luglio 2009 0.28
To: hobbit at hswn.dk
Subject: RE: [hobbit] dbcheck for RAC connections refused
Hi !
Unfortunatley, i've got no luck till yet ...
I've cleaned up my dbcheck.ini like this, also without any spaces ...
cat etc/dbcheck.ini|grep -v \#
[default]
warningout              = send
conn                    = conn
include                 =
defaultgroup            =
timeoutdbconn           = 20s
timeoudbtselect         = 1h
oraclehome              = /opt/products/oracle-client/10.2g/
fastcheck               = 30m
mediumcheck             = 2h
slowcheck               = 6h
bbdisplay               = localhost
bbsendtype              = multi
singlesendashost        = no
bbconncolumn            = ChkConn::fast
bbauditcolumn           = Audit::fast
bbtablespacecolumn      = TblSpace::fast
bbextentcolumn          = Extent::slow
bbhitcachecolumn        = HitCache::fast
bbmemreqcolumn          = MemReq::fast
bbrollbackcolumn        = RollBack::slow
bbinvobjcolumn          = InvObj::fast
bblockscolumn           = Locks::fast
bbsessioncolumn         = Session::fast
bbsessionuserlist       = yes
username                = XXXXXXXXXX
password                = XXXXXXXXXXXXXXxx
tblspcheck              = 90:95
tblspreportautoext      = 90:95
pinlibcheck             = 95:90
blbuffcheck             = 95:90
hitcachecheck           = 95:90
rowcachecheck           = 95:90
extentcheck             = 90:95
rollbackcheck           = 90:95
sessioncheck            = 90:95
processcheck            = 90:95
rb.SYSTEM               = 101:101
invobjwarn              = :::
invobjignore            =
dbtype                  = Oracle
[atlast1]
hostname                = lcg3d-a-v-1
port                    = 1521
username                = XXXXXXXXXXX
password                = XXXXXXXXXXXXX
bbauditcolumn           = Audit:notest
bbtablespacecolumn      = TblSpace:notest
bbextentcolumn          = Extent:notest
bbrollbackcolumn        = RollBack:notest
bblockscolumn           = Locks:notest
bbinvobjcolumn          = InvObj:notest
[atlast2]
hostname                = lcg3d-a-v-2
port                    = 1521
bbauditcolumn             = Audit:notest
bbtablespacecolumn      = TblSpace:notest
bbextentcolumn          = Extent:notest
bbrollbackcolumn        = RollBack:notest
bblockscolumn           = Locks:notest
bbinvobjcolumn          = InvObj:notest
username                = XXXXXXXXXXx
password                = XXXXXXXXXXXXXX
[atlast3]
hostname                = lcg3d-a-v-3
port                    = 1521
bbauditcolumn             = Audit:notest
bbtablespacecolumn      = TblSpace:notest
bbextentcolumn          = Extent:notest
bbrollbackcolumn        = RollBack:notest
bblockscolumn           = Locks:notest
bbinvobjcolumn          = InvObj:notest
username                = XXXXXX
password                = XXXXXXXXXX
[atlast4]
hostname                = lcg3d-a-v-4
port                    = 1521
username                = XXXXXXX
password                = XXXXXXXXXXX
bbauditcolumn             = Audit:notest
bbtablespacecolumn      = TblSpace:notest
bbextentcolumn          = Extent:notest
bbrollbackcolumn        = RollBack:notest
bblockscolumn           = Locks:notest
bbinvobjcolumn          = InvObj:notest
[atlas_tag_db]
oraclerac               = yes
port                    = 1521
username                = XXXXXXXX
password                = XXXXXXXXX
bbhitcachecolumn        = HitCache:notest
bbmemreqcolumn          = MemReq:notest
The output of the test is :
l Stats
Test Time    = 00:00:01
Host Checked = 5
Status Msg   = 29
Colors                   Colors
clear                       24
red                          5
Events                   Events
Audit                        1
ChkConn                      5
Extent                       1
HitCache                     4
InvObj                       1
Locks                        1
MemReq                       4
RollBack                     1
Session                      5
TblSpace                     1
uptime                       5
Types                     Types
status+30m                  22
status+60                    5
status+6h                    2
Hosts Summary
Hosts                     clear        red     Number      Times
atlas_tag_db                  8          1          9   00:00:00
atlast1                       4          1          5   00:00:00
atlast2                       4          1          5   00:00:00
atlast3                       4          1          5   00:00:00
atlast4                       4          1          5   00:00:00
Hosts Summary
Hosts                     Audit    ChkConn     Extent   HitCache
InvObj      Locks     MemReq   RollBack    Session   TblSpace     uptime
atlas_tag_db                  1          1          1          0
1          1          0          1          1          1          1
atlast1                       0          1          0          1
0          0          1          0          1          0          1
atlast2                       0          1          0          1
0          0          1          0          1          0          1
atlast3                       0          1          0          1
0          0          1          0          1          0          1
atlast4                       0          1          0          1
0          0          1          0          1          0          1
Errors
No General Errors
atlas_tag_db
Fri Jul 31 00:02:15 2009:ERROR: Timeout connecting to
DBI:Oracle:atlas_tag_db!
atlast1
Fri Jul 31 00:02:15 2009:ERROR: Timeout connecting to
DBI:Oracle:host=lcg3d-a-v-1;port=1521;sid=atlast1!
atlast2
Fri Jul 31 00:02:15 2009:ERROR: Timeout connecting to
DBI:Oracle:host=lcg3d-a-v-2;port=1521;sid=atlast2!
atlast3
Fri Jul 31 00:02:15 2009:ERROR: Timeout connecting to
DBI:Oracle:host=lcg3d-a-v-3;port=1521;sid=atlast3!
atlast4
Fri Jul 31 00:02:15 2009:ERROR: Timeout connecting to
DBI:Oracle:host=lcg3d-a-v-4;port=1521;sid=atlast4!
Warnings
No General Warnings
...
In tnsnames.ora (under ORACLE_HOME/network/admin) I've those
configuration:
This is a good point, i can't use the tnsnames.ora, because my
oracle-software is placed in the global filesystem AFS ...
So, could this be a problem, if i have to got all informations only
from  dbcheck.ini ?
...
You can also try to connect with a telnet to the hostname and
listener  port
Yep, telnet to port 1521 is working .. to  lcg3d-a-v-1/4
here my enviroment:
bin/bbcmd  ext/dbcheck.pl -v
2009-07-31 00:04:28 Using default environment file
/usr/lib/hobbit/server/etc/hobbitserver.cfg
Program Information
dbcheck.pl version 1.09
Hobbit_fd_lib.pm version 1.16
OS Information
Linux it-wgs02 2.6.9-78.0.22.ELsmp #1 SMP Thu Apr 30
23:30:54 CDT
2009 i686 i686 i386 GNU/Linux
Perl version is v5.8.5
Module Information
FindBin 1.44
POSIX 1.08
Text::Wrap 2001.09292
Time::Local 1.1
Sys::SigAction 0.04
Config::IniFiles 2.39
DBI 1.605
DBD:: DBM 0.03
DBD:: ExampleP 12.010007
DBD:: File 0.35
DBD:: Gofer 0.011377
DBD:: Oracle 1.23
DBD:: Proxy 0.2004
DBD:: Sponge 12.010002
Hmm, misterious ..
    martin
On Thu, 30 Jul 2009, Francesco Duranti wrote:
...
This is what i normally do to check a RAC DB + single RAC instance
in  the configuration:
oraclehome              = /usr/local/oracle/10.2.0/client/
bbdisplay               = localhost
bbconncolumn            = ChkConn::fast
bbauditcolumn           = Audit::fast
bbtablespacecolumn      = TblSpace::fast
bbextentcolumn          = Extent::slow
bbhitcachecolumn        = HitCache::fast
bbmemreqcolumn          = MemReq::fast
bbrollbackcolumn        = RollBack::slow
bbinvobjcolumn          = InvObj::fast
bblockscolumn           = Locks::fast
bbsessioncolumn         = Session::fast
bbsessionuserlist       = yes
username                = userxxx
password                = pwxxxxx
[CL10G]
oraclerac               = yes
bbhitcachecolumn        = HitCache:notest
bbmemreqcolumn          = MemReq:notest
[CL10G1]
hostname                = itromdb21
bbauditcolumn           = Audit:notest
bbtablespacecolumn      = TblSpace:notest
bbextentcolumn          = Extent:notest
bbrollbackcolumn        = RollBack:notest
bblockscolumn           = Locks:notest
bbinvobjcolumn          = InvObj:notest
[CL10G2]
hostname                = itromdb22
bbauditcolumn           = Audit:notest
bbtablespacecolumn      = TblSpace:notest
bbextentcolumn          = Extent:notest
bbrollbackcolumn        = RollBack:notest
bblockscolumn           = Locks:notest
bbinvobjcolumn          = InvObj:notest
In tnsnames.ora (under ORACLE_HOME/network/admin) I've those
configuration:
CL10G =
(DESCRIPTION =
(ADDRESS = (PROTOCOL = TCP)(HOST = ITROMDB21)(PORT = 1521))
(ADDRESS = (PROTOCOL = TCP)(HOST = ITROMDB22)(PORT = 1521))
(LOAD_BALANCE = no)
(CONNECT_DATA =
(SERVER = DEDICATED)
(SERVICE_NAME = cl10g)
)
)
CL10G2 =
(DESCRIPTION =
(ADDRESS = (PROTOCOL = TCP)(HOST = ITROMDB22)(PORT = 1521))
(CONNECT_DATA =
(SERVER = DEDICATED)
(SERVICE_NAME = cl10g)
(INSTANCE_NAME = cl10g2)
)
)
CL10G1 =
(DESCRIPTION =
(ADDRESS = (PROTOCOL = TCP)(HOST = ITROMDB21)(PORT = 1521))
(CONNECT_DATA =
(SERVER = DEDICATED)
(SERVICE_NAME = cl10g)
(INSTANCE_NAME = cl10g1)
)
)
With this setting I normally use the tnsnames.ora configuration
for RAC  (I was able to check RAC database only using tnsnames.ora not directly.
You  can try to remove the hostname and port from the db
configuration  for the RAC.
If the SID is equal to the name of the DB to check (the one in [])
you  can also remove the sid line from the configuration.
You can also try to connect with a telnet to the hostname and
listener  port to see if it answer (or better you can configure on
the client the  tnsnames.ora and do a tnsping to see if the instances are answering).
As you can see I don't do some checks on the RAC itselft or on the
single instances... For example I don't test memreq and hitcache on
the  RAC and I don't check TblSpace, Extent, Rollback, locks,
invalid objects  on the single instances.
Please check also on the configuration file and remove all ending
space  after host or port or sid. I don't know if it can create
problems (I'll  check this also and remove those from the
configuration option for the  next version).
Let me know if this work for you or you have other problems.
Francesco
-----Original Message-----
From: Martin Flemming [mailto:martin.flemming at desy.de]
Sent: giovedì 30 luglio 2009 11.47
To: hobbit at hswn.dk
Subject: RE: [hobbit] dbcheck for RAC connections refused
Hi, Franceso !
Thanks for the hint, it looks better/cleaner now, but the error is
still  exist ...
"
/usr/lib/hobbit/server #  bin/bbcmd ext/dbcheck.pl -t atlast1
2009-07-30 08:08:28 Using default environment file
/usr/lib/hobbit/server/etc/hobbitserver.cfg
Thu Jul 30 08:08:29 2009:ERROR on atlast1: Timeout connecting to
DBI:Oracle:host=lcg3d-a-v-1   ;port=1521          ;sid=atlast1      !
"
[it-wgs02] /usr/lib/hobbit/server #  bin/bbcmd perl -d
ext/dbcheck.pl  -t atlast1  2009-07-30 08:05:59 Using default
environment file  /usr/lib/hobbit/server/etc/hobbitserver.cfg
Loading DB routines from perl5db.pl version 1.27  Editor support
available.
Enter h or h h' for help, or man perldebug' for more help.
Hobbit_fd_lib::CODE(0x9505830)(/usr/lib/hobbit/server/ext/Hobbit_fd_lib.pm:1222):1222:
&load_config ;  DB<1> n
Hobbit_fd_lib::CODE(0x9505830)(/usr/lib/hobbit/server/ext/Hobbit_fd_lib.pm:1223):1223:
&get_disabled_hosts;  DB<1> n
Hobbit_fd_lib::CODE(0x9505830)(/usr/lib/hobbit/server/ext/Hobbit_fd_lib.pm:1224):1224:
my $pv=sprintf "%vd", $^V;  DB<1> n
Hobbit_fd_lib::CODE(0x9505830)(/usr/lib/hobbit/server/ext/Hobbit_fd_lib.pm:1225):1225:
my $sv=$Sys::SigAction::VERSION;  DB<1> n
Hobbit_fd_lib::CODE(0x9505830)(/usr/lib/hobbit/server/ext/Hobbit_fd_lib.pm:1226):1226:
if (($pv =~ /5\.8\.0/) && ($sv !~ /0\.09/)) {  DB<1> n
main:: (ext/dbcheck.pl:39):      my ($orahome, @listahost, $dbhost,
main:: $port, $username, $password, $dbtype,
$sid,$onlycustomcheck);
main:: DB<1> n
main:: (ext/dbcheck.pl:40):      my
main::
(%warnhash,%alerthash,$invobjignore,$invobjwarn, at invobjwlist, at invob
jilist,$checkuserlist,%customcheck);
main:: DB<1> n
main:: (ext/dbcheck.pl:41):      my
main:: ($sendtv,$singlesendas,%selcmd,$raccheck,
main:: @KeyToCheck, at CustomKeyToCheck);  DB<1> n
main:: (ext/dbcheck.pl:43):      my %DBCHECK =
main:: (main::(ext/dbcheck.pl:44):                     oracle
main:: => [
main::
"fulltest","conn","audit","tablespace","extent","hitcache","memreq"
,
main:: (ext/dbcheck.pl:45):
main:: "rollback","invobj","locks","session" ],
main:: (ext/dbcheck.pl:46):                     oracledef        => [
main::
"Database","ChkConn::fast","Audit::fast","TblSpace::fast","Extent::
slow",
main:: (ext/dbcheck.pl:47):
main:: "HitCache::fast","MemReq::fast",
"RollBack::slow","InvObj::fast",
main:: (ext/dbcheck.pl:48):
main:: "Locks::fast","Session::fast" ],
main:: (ext/dbcheck.pl:49):                     informix         => [
main:: "fulltest","conn","tablespace" ],
main:: (ext/dbcheck.pl:50):                     informixdef      => [
main:: "Database","ChkConn::fast","TblSpace::fast" ],
main:: (ext/dbcheck.pl:51):                     sqlserver        => [
main:: "fulltest","conn","tablespace" ],
main:: (ext/dbcheck.pl:52):                     sqlserverdef     => [
main:: "Database","ChkConn::fast","TblSpace::fast" ],
main:: (ext/dbcheck.pl:53):                     mysql            => [
main:: "fulltest","conn","mysqlperf" ],
main:: (ext/dbcheck.pl:54):                     mysqldef         => [
main:: "MySQL","ChkConn::fast","mysqlperf::fast" ],
main:: (ext/dbcheck.pl:55):                     db2              => [
main:: "fulltest","conn","tablespace" ],
main:: (ext/dbcheck.pl:56):                     db2def           => [
main:: "DB2","ChkConn::fast","TblSpace::fast" ],
main:: (ext/dbcheck.pl:57):                   );  DB<1> n
main:: (ext/dbcheck.pl:59):      my $CHECKLIST = [ "pinlib:95:90",
main:: "blbuff:95:90", "rowcache:95:90", "tblsp:90:95",
"hitcache:95:90",
main:: "rollback:90:95",
main:: (ext/dbcheck.pl:60):
main:: "session:90:95", "process:90:95", "memreq:90:95", "extent:90:95"
main:: ];  DB<1> n
main:: (ext/dbcheck.pl:63):
main:: @listahost=&Hobbit_fd_lib::get_server_to_test($TEST) if
main:: (!@listahost);  DB<1> n
main:: (ext/dbcheck.pl:64):
main:: &Hobbit_fd_lib::execute_thread(\&execute_db_test
, at listahost);
main:: DB<1> n
Thu Jul 30 08:06:27 2009:ERROR on atlast1: Timeout connecting to
DBI:Oracle:host=lcg3d-a-v-1   ;port=1521          ;sid=atlast1      !
Debugged program terminated.  Use q to quit or R to restart,
use O inhibit_exit to avoid stopping after program termination,
h q, h R or h O to get additional info.
main::(ext/dbcheck.pl:65):              exit;
DB<1>
... and the logfile e.g.
Thu Jul 30 11:45:37 2009:ERROR on atlas_tag_db: Timeout connecting to
DBI:Oracle:host=lcg3d-a-v-4   ;port=1521         ;sid=atlas_tag_db  !
Thu Jul 30 11:45:37 2009:ERROR on atlast1: Timeout connecting to
DBI:Oracle:host=lcg3d-a-v-1   ;port=1521          ;sid=atlast1      !
Thu Jul 30 11:45:37 2009:ERROR on atlast3: Timeout connecting to
DBI:Oracle:host=lcg3d-a-v-3  ;port=1521          ;sid=atlast3       !
Thu Jul 30 11:45:37 2009:ERROR on atlast2: Timeout connecting to
DBI:Oracle:host=lcg3d-a-v-2   ;port=1521          ;sid=atlast2       !
Thu Jul 30 11:45:37 2009:ERROR on atlast4: Timeout connecting to
DBI:Oracle:host=lcg3d-a-v-4  ;port=1521         ;sid=atlast4   !
Any suggestions for further debugging ?
thanks & cheers
   martin
On Thu, 30 Jul 2009, Francesco Duranti wrote:
...
Hi,
dbcheck.pl (and the other scripts in the package) use the
CONFIG::IniFiles to read configuration data. CONFIG::Inifiles
allow  comment line starting with "#" and ";" but I think it will
not  recognize the "#" comment at the middle of a line so if you
need  comments you should put them alone on the line before or
next to the  line to comment on.
On some of the sample config file I've put some comment next to
parameters and this could probably generate confusion. I'll change
the  sample config file and put this behavious also in the
documentation  I'm writing for the next version (the actual
documentation is not  really well written and organized).
Francesco
On Wed, July 29, 2009 15:29, martin.flemming at desy.de wrote:
...
Hi !
I want to use dbcheck.pl from
http://sourceforge.net/projects/hobbit-perl-cl/
.. but unfortunatley something is going wrong, i get only
"Connection Refused" Messages and no data  .. :-(
My config is below:
[SNIP]
...
dbcheck.ini:
oraclehome              = /opt/products/oracle-client/10.2g/
username                = XXXX
password                = XXXXXXXXXXXXXXXXXXXXXXX
[atlas_tag_db]
dbtype                  = Oracle
oraclerac               = yes
hostname                = lcg3d-a-v-4   # hostname of the rac
instance n.2
sid                     = atlas_tag_db  # sid of the rac instance
n.2
port                    = 1521          # port of listener on rac
instance
I don't use this code, but your error output suggests that it's
trying  to  use your hostname, sid, and port from dbcheck.ini
exactly as written,  comments and all.  I don't see anything in
the sample dbcheck.ini file  that suggests that you can use
trailing comments in the way that you  are  doing, so perhaps you
should try without those.
To unsubscribe from the hobbit list, send an e-mail to
hobbit-unsubscribe at hswn.dk
To unsubscribe from the hobbit list, send an e-mail to
hobbit-unsubscribe at hswn.dk

[hobbit] dbcheck for RAC connections refused

fduranti＠q8.it

cat etc/dbcheck.ini|grep -v \#