Options Reference
In GSCC, there are several configuration files in the /etc/gscc
which define the cluster:
gscc.conf
<member>.def
clipup
heartbeat
gsuidcustom.conf
The available options will be described in the following sections.
gscc.conf
gscc.conf lets you customize the general behaviour of gscc on the system, independent of individual cluster members. The most used options in gscc.conf are:
FileLogLevel
The FileLogLevel option describes the verbosity of logging written in /var/gscc/logs
Syntax:
FileLogLevel <Value>
- Values:
Debug
- Debug output, Behavioural information, Noticable events, Errors and Warnings will be logged.Info
- Behavioural information, Noticable events, Errors and Warnings will be logged.Notice
- Noticable events, Errors and Warnings will be logged.Warning
- Errors and Warnings will be logged.Error
- Only GSCC errors will be logged.
SysLogLevel
Syntax:
SysLogLevel <Value>
Values: The same as for FileLogLevel: Debug
, Info
, Notice
, Warning
and Error
Port
The Port option specifies the port number for the Cluster communication daemon (gscomd). If not specified, the default port will be used: 3939. This option is mainly used when multiple gscomd daemons have to run on the same physical host (like having multiple GSCC instances in separate containers).
Syntax:
Port <Value>
Environment Variables
GSCC environment variables define the cluster behaviour in several situations.
Environment GratuitousARP
This option (if GratuitousARP=yes) will cause GSCC to send Gratuitous ARP when addind an IP alias. The tool used behind is arping (in Unsolicited ARP mode, to update neighbours). Sending Gratuitous ARP can avoid isues like clients behind a netwrok gateways not being able to access the ISP Server IP address after a takeover The gateway might not have update the arp cache). The outpot of the arping command (and the command itslef) can be seen in GSCC Action Log every time when am_addalias is used (ISP Server start, restart, instance takeover or using the Operator command to add the IP alias).
Syntax:
Environment GratuitousARP=<Value>
Values: yes
or no
(default: no)
Example:
[gscc-S33V1 root@m2 logs]$ cat /etc/gscc/gscc.conf |grep -i arp
Environment GratuitousARP=yes
GSCC Action log:
221125100447 m2 S33V1 am_addalias(29132) Method started.
221125100447 m2 S33V1 am_addalias(29132) INFO: Adding IP alias addresses.
221125100447 m2 S33V1 am_addalias(29132) INFO: ip addr add dev ens33 local 192.168.178.215/255.255.255.0 broadcast 192.168.178.255 label ens33:S33V1
221125100447 m2 S33V1 am_addalias(29132) INFO: 192.168.178.215 added successfully!
221125100447 m2 S33V1 am_addalias(29132) INFO: Sending Gratuitous ARP: /usr/sbin/arping -U 192.168.178.215 -I ens33 -c 3
221125100449 m2 S33V1 am_addalias(29132) ARPING 192.168.178.215 from 192.168.178.215 ens33
221125100449 m2 S33V1 am_addalias(29132) Sent 3 probes (3 broadcast(s))
221125100449 m2 S33V1 am_addalias(29132) Received 0 response(s)
221125100449 m2 S33V1 am_addalias(29132) INFO: Adding IP addresses successful.
221125100449 m2 S33V1 am_addalias(29132) Method completed.
Note that Received 0 response(s)
is normal, as arping is used in Unsolicited mode (no answer expected).
Environment ISPConsole
Syntax:
Environment ISPConsole=<Value>
Values: yes
or no
(default: no)
If set to yes, this option activates the ISP console at startup for all Members. It’s possible to activate this option only for a specific Member - in this case we have to use tsm_server_access
option with value console
in Member.def.
Other related options:
Environment ISPConsoleAlwaysOn=<Value> - Keep writing to the console log file even after startup is complete (default: no)
Environment ISPConsoleActionLoge=<Value> - Write Console output also into action method log (default: yes)
Environment ISPConsolePath=<Value> - Path where Console log is written (default: ISP home dir)
The output will be visible in the Action Logs, starting with ISP:
, followed by the console message.
Example:
[gscc-S33V1 root@m2 logs]$ grep ISPConsole /etc/gscc/gscc.conf
Environment ISPConsole=on
Action Logs:
221113123824 m2 S33V1 am_starttsm(25016) INFO: Waiting for ISP server to start.
221113123829 m2 S33V1 am_starttsm(25016) ISP: Version 8, Release 1, Level 15.000
221113123830 m2 S33V1 am_starttsm(25016) ISP:
221113123831 m2 S33V1 am_starttsm(25016) ISP: Licensed Materials - Property of IBM
221113123832 m2 S33V1 am_starttsm(25016) ISP:
221113123833 m2 S33V1 am_starttsm(25016) ISP: (C) Copyright IBM Corporation 1990, 2022.
221113123834 m2 S33V1 am_starttsm(25016) ISP: All rights reserved.
221113123835 m2 S33V1 am_starttsm(25016) ISP: U.S. Government Users Restricted Rights - Use, duplication or disclosure
221113123836 m2 S33V1 am_starttsm(25016) ISP: restricted by GSA ADP Schedule Contract with IBM Corporation.
221113123837 m2 S33V1 am_starttsm(25016) ISP:
221113123838 m2 S33V1 am_starttsm(25016) ISP: ANR7801I Subsystem process ID is 25977.
221113123840 m2 S33V1 am_starttsm(25016) ISP: ANR0900I Processing options file /SPORD/ISP/S33V1/config/dsmserv.opt.
...etc
221113123931 m2 S33V1 am_starttsm(25016) ISP: ANR0993I Server initialization complete.
221113123932 m2 S33V1 am_starttsm(25016) ISP: IBM Spectrum Protect:S33V1>
221113123933 m2 S33V1 am_starttsm(25016) ISP: ANR0916I IBM Spectrum Protect distributed by International Business Machines is
221113123934 m2 S33V1 am_starttsm(25016) ISP: now ready to use.
221113123934 m2 S33V1 am_starttsm(25016) INFO: ISP server ready.
221113123934 m2 S33V1 am_starttsm(25016) INFO: Terminating initial ISP Console Logging.
By using this option, any eventual errror during ISP satrt will be visible exactly as dsmserv would have been started in foreground.
By default, the console loggind will be stopped after Server initialization. If desired, the logging can be keept on by using Environment ISPConsoleAlwaysOn=yes
.
Environment Docker
Syntax:
Environment Docker=<Value>
Values: yes
or no
(default: no)
This option, if set to yes, enables GSCC to manage ISP instances configured inside a contianer (Docker, Podman). That’s the only mandatory option required by GSCC to manage containerized instances.
Note that this option works in realation with SPORD (Spectrum Protect on Resilient Docker), which is a separate package, providing a CLI to deploy and manage containerized ISP instances.
If multiple instances are running on the same host, each Port
value from gscc.conf for every instance has to be unique and same thing is true for the gsuid port specfied in gsuidcustom.conf (the port use for the Web server, aloowing GSCCAD to work).
Environment DockerMon
Syntax:
Environment DockerMon=<Value>
Values: yes
or no
(default: no)
This option works only in SPORD environamnets (when Environment Docker=on
) and allows GSCC to use a Monitoring container, acting like a 3rd party by checking the status of the ISP containers and take spefic actions, when required. Even if this option isn’t mandatory, it can be usefull in situations whe the ISP contaieners crashes often, while the physical system stays on. The Monitoring container will help in performning takeovers in such situations.
Environment TSMcfgSyncAlways
Syntax:
Environment TSMcfgSyncAlways=<Value>
Values: yes
or no
(default: no)
When a cluster is joined, ISP configuration files (volhist, devconf, dsmserv.opt) and also the security files (cert.rdb, cert.crl, cert.sth, cert.kdb, cert256.arm, dsmkeydb.kdb, dsmkeydb.sth) are synced to the remote side, if contains changes. By specyfing this option with the value yes
we can force the sync of the security files even if these file are unchanged. There is a dedicated log entry for syncing in /var/gscc/logs/MEMBER.sync.
Related option:
Environment TSMcfgLegacy=<Value> (default: no)
The legacy file sync (dsmserv.pwed, cert.arm) for ISP instances before 7.1.2 can be enabled if this option is specified with the value yes
.
Environment Autoreintegrate
Syntax:
Environment Autoreintegrate=<Value>
Values: yes
or no
(default: yes)
After a HADR takeover ISP will be runnig as a Standard DB (HADR stopped). If the previous Primary becomes again available and the cluster is Joined, GSCC will automatically attempt an Reintegration. As consequence, the previous Primary will be started as a Standby and the current active Standard (where ISP is running) will be started as a Primary. All these actions are logged in Action Logs.
Setting the value to no
will dsiable this behaviour. Note that the Reintegration is also available as a manual tool (GSCCAD and GSCC cli) and it’s described in our documentation.
Environment delaytakeoverreaction
Syntax:
Environment delaytakeoverreaction=<Value>
Values: yes
or no
(default: no)
This parameter allows to delay a takeover reaction, if desired so. If the value is set to yes
, no takeover will occur for the first 10 minutes, in case Primary isn’t available. After the timeout (600 sec), if the Primary is still not reachable, the takeover will finally take place. There would be no actions (takeover) if the Primary will become available.
To delay a takeover using a custom value for the timeout use this Environment parametter in addidion:
Environment delaytakeoverreactionsec=<Value> (Value=seconds)
Please note, for using this parameter the HADR PEER_WINDOW_END value of DB2 instances have to be extended accordingly.
Environment UmountStgRetries
Syntax:
Environment UmountStgRetries=<Value>
Values: Any positive integer (default: 2)
This parameter determines how often GSCC tries to unmount shared storage when stopping/killing cluster resources on a node or when it is instructed to manually do so.
The value for Environment UmountStgRetryInterval
defines how long GSCC waits in between each Retry.
Environment UmountStgRetryInterval
Syntax:
Environment UmountStgRetryInterval=<Value>
Values: Any amount of seconds (default: 10)
This parameter sets the waiting time in seconds in between each attempt to unmount shared storage resources when stopping/killing cluster resources on a node or when it is instructed to manually do so.