Archiver Process on UNIX
The archiver process (dmfacp) does not start unless the recovery (dmfrcp) process is running
. However, an installation runs without an archiver process until the log file fills up. User programs are suspended as outstanding transactions in the log file are backed out
. For information about the
Ingres recovery state, see
Recovery Process Monitoring.
• Archiver process (dmfacp) startup errors are likely to result from:
• Improper shared memory resources
• Inability to read the transaction log file
• Inability to write journal files
Check for Archiver Process Errors on UNIX
Use the following procedure to check archiver process startup problems. Some of these checks are the same as for the recovery process:
1. Check that the shared memory resources are properly installed: Use the csreport utility and check that the size, ownership and permissions of the semaphore and shared memory segments meet the minimum requirements for your port. For requirements, see the Readme file.
The command to allocate Ingres shared memory and semaphores (when logged is as the installation owner) is csinstall. You can display them from UNIX with the command ipcs.
2. Make the following checks on the transaction log file “ingres_log”:
a. Verify that “ingres_log” exists at that location by entering the following command at the operating system prompt:
ls -l
If it does not exist, you must create it by running the ingbuild program.
b. Verify that “ingres_log” is owned by the user that owns the installation.
If not, issue the following command at the operating system prompt, where userid is the user who owns the installation:
chown userid ingres_log
c. If the transaction log file was created as:
An ordinary UNIX file, make sure it has permissions 660 (that is,
“-rw-rw----”). If not, issue the following command at the operating system prompt:
chmod 660 ingres_log
A raw log, permissions is “crw------”.
Check that the II_JOURNAL location is a valid location. Issue the following command at the operating system prompt:
infodb | grep ii_journal
3. Check journal locations by verifying that:
a. The journal location name points to a valid directory containing subdirectories “ingres/jnl/default/dbname”
b. The permissions on these directories are:
- 755 for the ingres directory
- 777 for the default directory
4. Check that the disk partition containing the journal files is not 100% full. Issue the following command at the operating system prompt:
df
If a journal partition is 100% full, it is impossible to write journals and the Archiver stalls. If this is the reason preventing Archiver startup, you must either free space on the journal partition or temporarily disable journaling with the alterdb command.
5. If you corrected an Archiver process problem, verify that Ingres starts normally:
a. Shut down the partially started installation with the ingstop command.
b. Restart the installation with the ingstart command.
Last modified date: 08/29/2024