TOC PREV NEXT INDEX

Put your logo here!


4 TROUBLE SHOOTING GUIDE

This sections is intended to provide a first level debugging help. More can be found in the specific documentation. The following list is, of course, not exhaustive, and is made up with cases which occurred during previous installations. Please fell free to contribute to it (see 3.8). A case you have encountered and solved and that is not described hereafter, can occur again to somebody else. Let's help each other to save time!

4.1 Problems with NFS files

· NFS does not allow to mount a disk already mounted with NFS. Check that each NFS-mount is to the machine where the files are really located.

4.2 Missing setuid on MSQL deamon

SYMPTOM:

vltmgr:~ 41 > vccEnvStart -e <environment>

vccEnvStart@<host>: Warning: VccInfo: no valid result from query: GetByEnv <environment> envType

vccEnvStart@<host>: Warning: VccInfo: no valid result from query: GetByEnv <environment> hostName

....

RESPONSE:

ls -ld `which msqld`

-r-sr-x--- 1 vltmgr vlt 268607 Oct 6 22:05 msqld

If the permissions, owner and group don't match as above

then fix them.

REMARKS: if the permissions are correct and the message:

... Warning: VccInfo: no valid result from query...

is still there, it is probably due to some missing records in the accData.sql file. Check that all the entry in each table is appropriate and no information regarding your environment(s) are missing.

4.3 "couldn't write PID" after starting the mSQL server

Check the right ownerships of the file $VLTDATA/msql/msql.pid.

It should appear as follows:

-rw-r--r-- 1 vltmgr vlt 4 Dec 11 08:23 msqld.pid

4.4 The LCU does not boot

· check that the LCU is properly connected to the network
· on the LCU check the boot parameters (see 3.4.6): the host name, where to look for the VW boot file, the username
· on the WS check that: the "vx" username is defined and that the VxWorks file is readable by such a user.
· check that "vx" doesn't have a .cshrc that produces output!

4.5 The LCU boots but does not find the bootScript file

· on the LCU check that the bootScript file is correctly specified. Remember that NFS does not manage links: you shall specify the physical absolute path.
· on the WS check that:
· VLTROOT can be NFS mounted (see 2.3.3)
· the file is readable by the "vx" user (see 3.4.6)

4.6 LCC does not start successfully

· on the WS check that:
· VLTROOT can be NFS mounted
· bootScript is correct and contains the absolute path to VLTROOT and BOOTROOT
· the user specified in the bootScript exists on your UNIX system. To make it simple, use "vx" for both booting the LCU and in the bootScript file.
· logfile and rebootFile are available in $BOOTROOT/ENVIRONMENTS/<lcuenv>, even as empty files, and readable/writable by the "vx" user.
· devicesFile is available in the $BOOTROOT/ENVIRONMENTS/<lcuenv>. Remember that even if you do not have any devices, the file must exist and with a minimal content, see devicesFile(5) and being readable by the "vx" user.

4.7 VxWorks code is not compiled

· check VxWorks environment (see 2.7.3)

4.8 WS environment does not start

· check CCS environment (see 3.4.4)
· check that operating system parameters have been changed according to the CCSlite installation.
· if the logfile: $VLTDATA/ENVIRONMENTS/<env_name>/.ccsScheduler.log reports the following

+++++++++++++++++++++++++++++++++++++++++++++++++++++++++++

ccsScheduler: S-BIT not set for ccsScheduler1

+++++++++++++++++++++++++++++++++++++++++++++++++++++++++++

run the following commands:

chown root /vlt/<RELEASE>/CCSLite/bin/ccsScheduler1

chmod u+s /vlt/<RELEASE>/CCSLite/bin/ccsScheduler1

4.9 Lccei does not start (properly)

· check that the PATH contains $VLTROOT/bin
· check that $VLTROOT/config/Lccei has been loaded by xrdb
· check that $VLTROOT/CDT contains valid files (the format can be checked using lcccdt), one for every process defined in $VLTDATA/ENVIRONMENTS/<lcuenv>/PROCESSES file
· check that qsemu is running

4.10 Lccei cannot communicate with an LCU

· check that qsemu is up and running
· check $VLTDATA/config/logLCU.config
· check that the environments defined on WS (qsemu -env) and LCU (lqsPrintEnvTbl) are consistent.

4.11 LOG system does not work

There are several facts that may affect the syslog, either at system level or due to bad VLT Software configuration. The following may help to find some of the possible causes:

· check that syslog is working:
$ ps -efa | grep syslog
· create a log in one of the system files (the commands used in the following example are for documentation only and may not work in your configuration. Please check them with your system manager):
· have a look in /etc/syslog.conf and chose a "selector"
· use logger to create an entry in the corresponding log file (the following is from an HP):
$ logger -p mail.debug "aaa"
· check that the file associated with the selector
$ tail /usr/spool/mqueue/syslog
Mar 06 17:14:44 te13 vltsccm: aaa
· check that the VLT_LOG_... variables are defined and that the content of /etc/syslog.conf is consistent with such a definition. Restart the syslogd to be sure that it is using the right files:
$ kill -HUP `cat /etc/syslog.pid`
Be aware of the different behaviour of HP and SUN:
on HP: if not existing, the files are created.
on SUN: if not existing, the syslog DOES NOT create them! Create empty files as described in 3.4.2.
· put a log in each of the log files:
$ logger -p local1.warning "this should go in logFile"
$ tail $VLTDATA/tmp/logFile
................
$ logger -p local2.warning "this should go in logAuto"
$ tail $VLTDATA/tmp/logAuto
................
· if you have CCS environment, create a log from the WS :
$ logUserData aa aa aa
$ tail $VLTDATA/tmp/logFile
................
· if you have an LCU, create a log from the LCU:
· check that $VLTDATA/config/logLCU.config exists as well as the pair <lcuenv><wsenv>. If needed edit the file.
· check that logManager is active. If not start it: $ logManager &
· force logManager to read the configuration file: $ logConfig
· on the LCU, give the following command to create a log
$ rlogin <lcu>
<lcuenv>-> logData "test", 101, "this is a log from an LCU"

Remember that logMonitor can be used to display the current VLT logs.

4.12 lccei does not accept CDT

CDT in NOV95 format or older are not valid any longer. Please edit them according to the new format as described in the LCC User Manual.

4.13 Scan system does not start on the LCU

LCU Database in FEB95 format or older are not valid any longer. Please change the LCU DB according to the new format.

4.14 Declarative conflicts during installation

The include files of this release cannot be mixed with previous ones. May be you have forgotten to save the old installation before starting this installation (see 2.6.2).

4.15 Force an error log from an LCU

The following procedure allows you to test the complete chain LCU-log system:

· Start "lccei"
· Select your environment and the process lccServer from the pull-down menus "Environment" and "Process".
· Send the commands "LOGSTRT" and "ERRSTRT" to your LCU.
· Select the process rdbServer.
· Send the command "DBGAINF :PARAMS:TABLES.full_table(-1)" or to any other table or vector in the database. This will produce the error log:
(mapMapValue) "lccERR_INV_RECEL : invalid number of records/ fields (number: number: -1 expected: )" (Warning 63) Help: -
· Now select the log window from the pull-down menue "File". The Log Window will pop up.
· Select "Show All Messages" or "Show Matching Messages" and "Logging" and "Error".This will display the error log in the log window.

4.16 Security error messages from Tk

Q.C4- X server insecure (must use xauth-style authorization) Tk requires you to have a secure X server before you can use the send command.

See the question 2.A.7 "How can I get Tk3.3 to even start - I get security error messages." in Thomas Accardo's Tk Toolkit Usage FAQ as well as http://ce-toolkit.crd.ge.com/tkxauth/ for instructions on how to make your server secure.

4.17 LCU boot problem: INTROOT contains old code

You may experience problem at boot time if you have INTROOT defined and old software is loaded there.

`lcuboot' will load modules automatically from INTROOT if they are there. If, for any reasons, old versions of LCC of of some other modules loaded during the boot are stored there, the boot process may be not complete or, if severe errors - like unresolved symbols - are found it may also abort.

After installation of any new version, current INTROOTs, if any, should be cleaned and the software regenerated using the newest version in VLTROOT.

4.18 LCU boot problem: tim board not installed

When LCC tries to read the time-board. If no board is installed, an error message is produced but the booting process continues regularly. If you do not have such a card, please ignore the message.

4.19 "unbalanced" Parentheses Warning from tclCheck

The "warnings" arise from the utility `tclCheck', which is automatically called by vltMake for every TCL file. Such "unbalanced" parentheses warning in correct code can be removed by using "\" as in the following examples:


> > File ccseiDb.tcl:
> > Inside a string: unmatched ( ending line 294 char 27
> >
> 294: set i [string first "(" $att]
E.g.: replace by: set i [string first "\(" $att]

> > File ccseiDbBrowser.tcl:
> > Expecting } got ] : line 309 char 31
> >
> 309: # -attrGeometry $attrGeometry]
also: # -attrGeometry $attrGeometry\]

Line 385: set line " ("
replace with: set line " \("

Line 415: append line " 1==1)"
replace with: append line " 1==1\)"

4.20 Different LCU environments on the same node

If there are several environments assigned to the same node and, by mistake, one writes two or more of them in the logLCU.config file, sometime errors occur (no reply from LCU, message "environment not active" etc.).

NEVER PLACE IN logLCU.config two or more LCU environments associated to the same node !!!

4.21 LCU slow speed.

During ASM commissioning, we pointed out a dramatic slow speed of the operations on a LCU: A process supposed to complete in 16 seconds was running more than 5 minutes. After investigation, we found out that a flag was set on the LCU, that enables run-time printouts on the LCU console, while no rlogin <lcu> had been issued. The effet is then that the serial line (9600 bps) is used to issue the approx. 300000 characters output during the sequence. A rough calculation of the output time is:

300000 char * 10 bits/char / 9600 bps > 300 seconds ( = 5 minutes)

Therefore, take care of this kind of setting (the flag was set in the userScript) that may terribly bias the performance of a LCU if no rlogin is used.


____oOo____



Quadralay Corporation
http://www.webworks.com
Voice: (512) 719-3399
Fax: (512) 719-3606
sales@webworks.com
TOC PREV NEXT INDEX