Creating a dump in Linux

First published: 16 December 2024 (cf, gs)

Information

This article explains the creation of a memory dump of Fabasoft services, including fetching the correct pid, in Linux.

Solution

Fetching the pid

Get PID for Fabasoft Folio Web services / Conversion services

ps ax | grep WebService

Because of Apache, Folio Web/Conversion services use two process id’s per service. The correct PID (tagged on the left) of the Web/Conversion service for a memory dump is the Sl type (tagged in the middle) service. With multiple services, the service number is contained in the path (100, 101, etc).

Get PID for COO/MMC Services, AT Services

Directly get the PID of the respective service, e.g.

• Using ps aux | grep <Servicename>
• fscmgmt list all

Create the dump

Run

gcore -a <pid>

The memory dump is saved in the current directory as core.<pid>, e.g. core.29991.

gcore may need to be installed for first usage.

Run Collect Dump Info

To collect all libraries to debug the memory dump, run the tool Collect Dump Info.