Common External Platform-Related Errors
This topic covers some of the errors reported by customers and their solutions.
OS initialization failed
Symptoms
Your platform agent does not startup properly on the remote UNIX system; in Redwood Server, the process server associated with the remote UNIX system remains in status connecting. Restarting the platform agent on the remote UNIX server appears to solve the issue in some cases, although the same behavior occurs again when the UNIX server is restarted.
Platform
This can happen on any UNIX/VMS platform.
Specific Error Message
This error message is displayed after debugging has been enabled:
/opt/redwood/agent/2023_3_0_20231116_10/bin/network-processor: OS initialization failed
debug 2023-11-16 06:07:45 CET [6706-platform-agent] Wait result = 5
error 2023-11-16 06:07:45 CET [6706-platform-agent] Scheduler network-processor instance 'prod' died with exitcode 5, trying to restart in 10 seconds...
debug 2023-11-16 06:07:45 CET [6706-platform-agent] Creating pidDir: /opt/redwood/agent/etc/pid/prod
debug 2023-11-16 06:07:45 CET [6706-platform-agent] Writing Parent PID to /opt/redwood/agent/etc/pid/prod/prod.ppid
info 2023-11-16 06:07:45 CET [6706-platform-agent] Starting network-processor instance 'prod'
debug 2023-11-16 06:07:45 CET [6706-platform-agent] /opt/redwood/agent/0_33_92/bin/network-processor -p /opt/redwood/agent/etc/pid/prod/prod.pid -i prod -n /opt/redwood/agent/net -l
Resolution
The environment for the platform agent is not correct; the platform agent does not have the correct character set in its environment and the correct character set cannot be guessed. The platform agent needs a character set that can be converted to UTF-8 using I18N tools. Ensure LC_ALL or, if it is not set, the LANG or LC_CTYPE environment variable is set to a compatible value for the agent process.
On Linux, ISO-8859-15, en_US.UTF-8, or de_DE.UTF-8 are compatible, for example. Note that LC_ALL takes precedence over LANG as well as all other LC_* variables.
opsys.socket - Connecting socket to localhost port 10180 failed: No connection
Symptoms
You are trying to retrieve a file with jgetfile and receive this error.
Platform
This can happen on all platforms which support jgetfile.
Specific Error Message
The following error messages can be seen in the terminal or log:
error 2023-11-16 06:07:45,124 Europe Standard Time [1696-jgetfile tid=0] opsys.socket - Connecting socket to localhost
port 10180 failed: No connection could be made because the target machine actively refused it.
error 2023-11-16 06:07:45,124 Europe Standard Time [1696-jgetfile tid=0] http.http - Unable to connect to localhost:10180
fatal 2023-11-16 06:07:45,124 Europe Standard Time [1696-jgetfile tid=0] jgetfile.main - Unable to process request
Resolution
Ensure the hostname in the registry entry /configuration/ContextURL is the FQDN of the central Redwood Server and not localhost or 127.0.0.1.
JCS-116012: Warning: Transient communication failure: HTTP response code: 500
Symptoms
Redwood Server was recently updated and the process servers are started for the first time. The platform agent is configured to update automatically, however, the update fails.
In the installation directory, which in this example is /opt/redwood/agent/, a new folder exists which is named 0_<m>_<n>, however, the security mode is not set correctly.
You may experience one or more of these symptoms, in which case you should perform the steps in resolution to find the exact issue.
In these examples, the following variables are used:
<m>- milestone, for example33.<n>- patch-level, for example12.<os>- operating system, for examplelinux.<platform>- the hardware platform, for examplex86./opt/redwood/agent- installation path.default- instance name of the platform agent.
Platform
This has been reported on GNU/Linux and it could affect other UNIX flavors.
Specific Error message
In Redwood Server, the following operator messages can be seen:
JCS-116012: Warning: Transient communication failure: HTTP response code: 500
JCS-116008: Platform agent update for "<os>-<platform>" to "M<m>.<n>" failed. The agent is still running version "M<m>.<n-1>". Aborting
Service "PlatformAgentService" on process server "<process_server>" stopped unexpectedly.
Resolution
The platform agent process must be restarted on the remote system with debug mode enabled to find the exact error; specify debug in /opt/redwood/agent/etc/startup/default/loglevel.
In this specific case, the updater could not update the following file, because it was installed as user root.
error 2023-11-16 06:07:45,739 CET [30968-http-request #7 tid=4147821424] opsys.file - Cannot open file '/opt/redwood/agent/etc/startup/default/version': Permission denied
error 2023-11-16 06:07:45,739 CET [30968-http-request #7 tid=4147821424] opsys.update - Cannot write to version file /opt/redwood/agent/etc/startup/default/version
$ ls -l /opt/redwood/agent/etc/startup/default/version
-rw-r--r-- 1 root root 7 Nov 10 13:41 /opt/redwood/agent/etc/startup/default/version
chown jdoe:jdoe /opt/redwood/agent/etc/startup/default/version
id: illegal option u
Platform
This error has been reported during the installation of the platform agent on Solaris and is due to the fact that the id command does not accept the POSIX-compliant -u switch.
Resolution
Temporarily put a POSIX-compliant id in the front of your path before executing the installation script.
On Solaris, a POSIX-compliant id binary can be found in /usr/xpg4/bin
$ PATH=/usr/xpg4/bin:$PATH
$ export PATH
$ chmod +x solaris-sparc.bin
$ ./solaris-sparc.bin
run_KSH: No such file or directory
Platform
This error has been reported on various UNIX flavors, among which GNU/Linux. This error occurs when the job-processor cannot locate ksh which must reside, or be symbolically linked to, /bin/ksh and needs to execute a KSH process.
Specific Error message
error 2023-11-16 06:07:45,811 WST [14188-Main tid=4158101184] opsys.job
- Cannot execute Redwood Job 81 '/home/user/jobs/81/run_KSH': No such file or directory
Resolution
Install ksh, pdksh or any ksh88 or ksh93 compatible shell.
onsiteTopic