When a process on the Steelhead fails, the operating system will generate a core dump of it: A copy of all the memory of the process as it was in use at the moment it failed. The size of the core dump depends on the size of the process, for a core dump of the optimization service this can be hundreds of megabytes.
After the core dump is generated, the process pm will tell the process mgmtd to generate a system dump and merge the two dumps into a process dump. This process might take a while, during which the process does not get restarted.
It can be necessary in the lifetime of a case that Riverbed TAC
requires a process dump of the optimization service. The way to
generate one is to use the command
debug service clone dump start
or the command
pm process <PROCESS> send-signal SIGQUIT
.
The command
debug service clone dump start
will create a clone of the optimization service first and generates
a process dump of the clone. This will cause the optimization and the
normal traffic to continue.
Figure 6.96. Generating a process dump of the process sport by cloning it
SH # debug service clone dump start [... some times passes ...] SH # show files process-dump SH-sport-20130102-123456.tar.gz SH # file process-dump SH-sport-20130102-123456.tar.gz ftp://192.168.1.1/incoming/SH-sport \ -20130102-123456.tar.gz SH # show info Current User: admin Status: Healthy Config: working Appliance Up Time: 12d 11h 20m 7s Service Up Time: 12d 11h 18m 14s Managed by CMC: yes Temperature (C): 44
The command
pm process <PROCESS> send-signal SIGQUIT
will terminate the optimization service and then generate
a process dump of it. This will cause the optimized TCP sessions to
be aborted and the optimization service to be restarted.
Figure 6.97. Generating a process dump of the process sport by restarting it
SH # pm process sport send-signal SIGQUIT [... some times passes ...] SH # show files process-dump SH-sport-20130102-123456.tar.gz SH # file process-dump SH-sport-20130102-123456.tar.gz ftp://192.168.1.1/incoming/SH-sport \ -20130102-123456.tar.gz SH # show info Current User: admin Status: Healthy Config: working Appliance Up Time: 12d 11h 24m 23s Service Up Time: 1m 31s Managed by CMC: yes Temperature (C): 44