When the number of queued update messages for some hosts reaches hundreds or thousands, there have been reports of a slowing down of all operations that need to use the clntd_s bridge on the Master, such as update of /etc/passwd, group files and some cadm or GUI operations.
A common cause is that there are many updates queued to hosts that cannot be contacted any longer. Perhaps the host has been decommissioned or will be down for a long time.
Resolution / Workaround
Start a BoKS prompt and run the following command:
boksdiag fque -bridge
This will show the IP address of the hosts and how many messages are queued for each one.
There are two ways of dealing with this.
The first approach is to use the command:
boksdiag fque -bridge -delete
and delete the queued messages for each host with, for example, more than a hundred stored messages.
The down side to this approach is that it needs to be done regularly and also if a host comes back up it will not automatically be updated, but the administrator will have to remember to bring it up to date using the command:
boksdiag updpsw -h
The second approach, and a recommended practice if the situation is becoming a problem, is to use BoKS' Host Pre-registration functionality to keep track of absent hosts and bring them up to date next time they appear in the domain. This would also delete any queued messages until the host is back.
This approach is described in article 11208 "How to configure an existing host in a BoKS Domain to use Host Preregistration functionality"
Still have questions? We can help. Submit a case to Technical Support.