Novell EDIRECTORY 8.8 SP3, Administration Manual, Page: 566 / 634

566 Novell eDirectory 8.8 Administration Guide
n
ov
do
cx (e
n)
11
Ju
ly 20
08
2b Select the "Hard Limit" radio button and set the 'Cache Maximum Size'.
2c Make sure to select the check box 'Cache Settings Permanent'.
3 Add preallocatecache=true to the _ndsdb.ini file.
The _ndsdb.ini is located in different locations, depending upon the operating system that
eDirectory is installed on.
Š
Netware - SYS:\_NETWARE\_NDSDB.INI.
NOTE: You can use edit.nlm to modify this file. Type the complete path name, when you
first start edit.nlm.
Š Windows - C:\novell\NDS\DIBFiles\_ndsdb.ini
Š
Linux, Solaris, and AIX - /var/nds/dib/_ndsdb.ini
NOTE: The last line of the ndsdb.ini file must be followed by a carriage return. This is
required for the Flaim to parse a file.
Verifying the Functioning of Preallocate
The amount of memory used should reflect at a minimum the cache amount set in the cache=
statement of the _ndsdb.ini file plus whatever memory has been allocated to current eDirectory
threads and its services since the last reboot.
Š
NetWare: Load Monitor. Go to System resources - Alloc Memory - Highlight DS.NLM. In
the screen above, you will see number in use. This will be at least the size specified in the
cache= statement in the _ndsdb.ini file.
Š
Linux: First find the process id: ps -e | grep ndsd. Then use top to find the amount of memory
allocated: top -p pid.
Š
Win32: Load TaskManager - Processes and observe the amount of memory used by dhost.
NOTE: Be careful while moving the roll forward logs. For more information on performing this
activity, refer to the TID 3668229 (http://www.novell.com/support/
search.do?cmd=displayKC&docType=kc&externalId=3668229&sliceId=SAL_Public&dialogID=5
6381416&stateId=0%200%2056387184)
.
18.7 Keeping eDirectory Healthy
The health of directory services is vital to any organization. Regular health checks using Novell
iMonitor will keep your directory running smoothly and will make upgrades and troubleshooting
much easier.
18.7.1 When to Perform Health Checks
In general, if your network doesn't change often (servers and partitions are added only every couple
of months and only simple changes are made frequently), perform health checks once a month.
If your network is more dynamic (partitions or servers are added weekly or your organization is
reorganizing), perform health checks weekly.