Raymond,
An FAILED_IMPORT handling upcall script just needs to consist of the
lconf --recover statement using the arguments passed after FAILED_IMPORT
on the command line appropriately.
You will need to add something to it for it to be able to determine the
means for reading the configuration information. You can use the shared
XML file with --select - as you have already being doing - or LDAP, or a
HTTP server - where you trigger the http served (XML) data to be updated
whenever you change the active server - to do this.
I would recommend having such a script source a configuration file that
is shared via NFS or replicated locally in /etc to obtain the details of
how to obtain the configuration information.
For example, something like the following for an LDAP setup:
#!/bin/bash
# source the config info
# defines LDAPURL and CFG_NAME
. /etc/lustre.cnf
# Get the Upcall Type
UPCALL_TYPE=3D"${1}"
shift
case "${UPCALL_TYPE}" in
FAILED_IMPORT)
lconf --recover \
--tgt_uuid ${1} \
--client_uuid ${2} \
--conn_uuid ${3} \
--ldapurl ${LDAPURL} \
--config ${CFG_NAME}
;;
esac
If you go the LDAP way you will need to setup an LDAP server (see the
Lustre Project Wiki) and also use the lactive command to update the
active server for an OST or MDS before starting it on the server.
Fergal.
--
Fergal.McCarthy@HP.com
(The contents of this message and any attachments to it are confidential
and may be legally privileged. If you have received this message in
error you should delete it from your system immediately and advise the
sender. To any recipient of this message within HP, unless otherwise
stated, you should consider this message and attachments as "HP
CONFIDENTIAL".)
-----Original Message-----
From: lustre-discuss-admin@lists.clusterfs.com
[mailto:lustre-discuss-admin@lists.clusterfs.com] On Behalf Of raymondyu
Sent: 09 March 2005 02:23
To: lustre-discuss@lists.clusterfs.com
Subject: [Lustre-discuss] Upcall script
Hi, I know that lustre can use ldap for failover. Any example or sample
of
upcall script? thanks
[uml27]$ lconf -v --reformat --ldapurl ldap://uml8 --config
config --timeout=3D100
--lustre_upcall=3D<your-path-tolustre-client-upcall>
_______________________________________________
Lustre-discuss mailing list
Lustre-discuss@lists.clusterfs.com
https://lists.clusterfs.com/mailman/listinfo/lustre-discuss