|
SLES /
Workaround for iSCSI issues on SLES11SLES.ISCSIOnSLES11 HistoryHide minor edits - Show changes to output November 13, 2009, at 04:07 PM
by -
Changed line 104 from:
for mountpoint in `grep "iscsi," /etc/fstab | sed -e 's|\s*iscsi,.*||' -e 's|.*/|/|'`; do to:
for mountpoint in `grep "iscsi," /etc/fstab | sed -e 's|\s*iscsi,.*||' -e 's|.*[ \t]/|/|'`; do Changed line 116 from:
for mountpoint in `grep "iscsi," /etc/fstab | sed -e 's|\s*iscsi,.*||' -e 's|.*/|/|'`; do to:
for mountpoint in `grep "iscsi," /etc/fstab | sed -e 's|\s*iscsi,.*||' -e 's|.*[ \t]/|/|'`; do Changed line 128 from:
for mountpoint in `grep "iscsi," /etc/fstab | sed -e 's|\s*iscsi,.*||' -e 's|.*/|/|'`; do to:
for mountpoint in `grep "iscsi," /etc/fstab | sed -e 's|\s*iscsi,.*||' -e 's|.*[ \t]/|/|'`; do November 10, 2009, at 01:59 PM
by -
Changed line 32 from:
UUID=81ff75b2-c311-46ea-b1b9-bade8b597624 /data iscsi,ext3 acl,user_xattr,sync,dirsync,nofail 1 2 to:
UUID=81ff75b2-c311-46ea-b1b9-bade8b597624 /data iscsi,ext3 acl,user_xattr,sync,dirsync,nofail,noatime 1 2 November 10, 2009, at 01:59 PM
by -
Changed line 32 from:
UUID=81ff75b2-c311-46ea-b1b9-bade8b597624 /data iscsi,ext3 acl,user_xattr,sync,dirsync,nofail,_netdev 1 2 to:
UUID=81ff75b2-c311-46ea-b1b9-bade8b597624 /data iscsi,ext3 acl,user_xattr,sync,dirsync,nofail 1 2 November 10, 2009, at 01:56 PM
by -
Changed line 18 from:
*Since the filesystem will not be available when the system first mounts the filesystems, it will fail when if reaches the iSCSI line, so you need to add the '''''nofail''''' option to the end of the option line. to:
*Since the filesystem will not be available when the system first mounts the filesystems, it will fail when if reaches the iSCSI line, so you need to add the '''''nofail''''' option to the end of the option line. (Note: on eariler systems nofail is not a vaild option so use '''''_netdev''''') November 10, 2009, at 11:50 AM
by -
Changed line 32 from:
UUID=81ff75b2-c311-46ea-b1b9-bade8b597624 /data iscsi,ext3 acl,user_xattr,sync,dirsync,nofail 1 2 to:
UUID=81ff75b2-c311-46ea-b1b9-bade8b597624 /data iscsi,ext3 acl,user_xattr,sync,dirsync,nofail,_netdev 1 2 Added lines 98-102:
#Added to take care of LVM groups on iSCSI vgscan > /dev/null sleep 1 vgchange -a y > /dev/null August 14, 2009, at 01:29 PM
by -
Changed line 99 from:
for mountpoint in `grep "iscsi," /etc/fstab | cut -f 2`; do to:
for mountpoint in `grep "iscsi," /etc/fstab | sed -e 's|\s*iscsi,.*||' -e 's|.*/|/|'`; do Changed line 111 from:
for mountpoint in `grep "iscsi," /etc/fstab | cut -f 2`; do to:
for mountpoint in `grep "iscsi," /etc/fstab | sed -e 's|\s*iscsi,.*||' -e 's|.*/|/|'`; do Changed line 123 from:
for mountpoint in `grep "iscsi," /etc/fstab | cut -f 2`; do to:
for mountpoint in `grep "iscsi," /etc/fstab | sed -e 's|\s*iscsi,.*||' -e 's|.*/|/|'`; do August 14, 2009, at 01:04 PM
by -
Changed lines 18-20 from:
*Since the filesystem will not be available when the system first mounts the filesystems, it will fail when if reaches the iSCSI line, so you need to add the nofail option to the end of the option line. *Since the filesystem is not physically attached, it is probably best to make sure that all writes are synchronus, so add the sync and dirsync options as well. to:
*Since the filesystem will not be available when the system first mounts the filesystems, it will fail when if reaches the iSCSI line, so you need to add the '''''nofail''''' option to the end of the option line. *Since the filesystem is not physically attached, it is probably best to make sure that all writes are synchronous, so add the '''''sync''''' and '''''dirsync''''' options as well. August 14, 2009, at 01:03 PM
by -
Changed line 17 from:
*As you can see in the example below, the filesystem type is '''scsi,''something'''''. Since there is not iscsi filesystem type it will mount the filesystem as '''''something''''' in this case '''''ext3'''''. to:
*As you can see in the example below, the filesystem type is '''scsi,''something'' '''. Since there is not a iscsi filesystem type it will mount the filesystem as '''''something''''' in this case '''''ext3'''''. August 14, 2009, at 01:02 PM
by -
Added line 11:
*Using the _netdev and auto options in the [@/etc/fstab@] file. Changed lines 16-17 from:
To make this script work it is necessary to slightly modify the iSCSI entries in the [@/etc/fstab@] file. As you can see in the example below, the filesystem type is '''scsi,''something''''' to:
To make this script work it is necessary to slightly modify the iSCSI entries in the [@/etc/fstab@] file: *As you can see in the example below, the filesystem type is '''scsi,''something'''''. Since there is not iscsi filesystem type it will mount the filesystem as '''''something''''' in this case '''''ext3'''''. *Since the filesystem will not be available when the system first mounts the filesystems, it will fail when if reaches the iSCSI line, so you need to add the nofail option to the end of the option line. *Since the filesystem is not physically attached, it is probably best to make sure that all writes are synchronus, so add the sync and dirsync options as well. August 14, 2009, at 12:57 PM
by -
Changed lines 15-16 from:
To make this script work it is necessary to slightly modify the iSCSI entries in the [@/etc/fstab@] file to:
To make this script work it is necessary to slightly modify the iSCSI entries in the [@/etc/fstab@] file. As you can see in the example below, the filesystem type is '''scsi,''something''''' August 14, 2009, at 12:56 PM
by -
Changed line 18 from:
(:div style="border-style:ridge; border-width:2px; background-color:#ffffcc; margin-left:50px; overflow:auto; width:650px; height:150px;":) to:
(:div style="border-style:ridge; border-width:2px; background-color:#ffffcc; margin-left:50px; overflow:auto; width:650px; height:175px;":) August 14, 2009, at 12:56 PM
by -
Changed line 18 from:
(:div style="border-style:ridge; border-width:2px; background-color:#ffffcc; margin-left:50px; overflow:auto; width:650px; height:300px;":) to:
(:div style="border-style:ridge; border-width:2px; background-color:#ffffcc; margin-left:50px; overflow:auto; width:650px; height:150px;":) August 14, 2009, at 12:54 PM
by -
Changed lines 20-28 from:
to:
/dev/system/swap swap swap defaults 0 0 /dev/system/root / ext3 acl,user_xattr 1 1 /dev/sda1 /boot ext3 acl,user_xattr 1 2 /dev/system/home /home ext3 acl,user_xattr 1 2 proc /proc proc defaults 0 0 sysfs /sys sysfs noauto 0 0 debugfs /sys/kernel/debug debugfs noauto 0 0 devpts /dev/pts devpts mode=0620,gid=5 0 0 UUID=81ff75b2-c311-46ea-b1b9-bade8b597624 /data iscsi,ext3 acl,user_xattr,sync,dirsync,nofail 1 2 Changed lines 35-134 from:
to:
#!/bin/bash # # /etc/init.d/open-iscsi # ### BEGIN INIT INFO # Provides: iscsi-mount # Required-Start: $network # Should-Start: iscsitarget # Required-Stop: $network # Should-Stop: iscsitarget # Default-Start: 3 5 # Default-Stop: # Short-Description: iSCSI initiator daemon wrapper # Description: The iSCSI initator is used to create and # manage iSCSI connections to a iSCSI Target. # This wrapper also takes care of iSCSI # fstab entries. # Created by Bob Brandt <projects@brandt.ie> # ### END INIT INFO ISCSISCRIPT=/etc/init.d/open-iscsi MOUNTSCRIPT=/etc/init.d/iscsi-mount RCSCRIPT=/sbin/rciscsi-mount declare -i overallstatus=0 test -h $RCSCRIPT || ln -sf $MOUNTSCRIPT $RCSCRIPT # Source LSB init functions . /etc/rc.status # Reset status of this service rc_reset iscsimount() { rc_reset echo -n "Mounting $1: " mount $1 rc_status -v return $? } iscsiumount() { rc_reset echo -n "Umounting $1: " umount $1 rc_status -v return $? } iscsicheck() { rc_reset echo -n "Verify if $1 is mounted: " mount | grep "on $1\b" > /dev/null rc_status -v return $? } iscsimountall() { # Find all fstab lines with iscsi as the fstype for mountpoint in `grep "iscsi," /etc/fstab | cut -f 2`; do # Only try to mount filesystems that are NOT currently mounted. if ! mount | grep "on $mountpoint\b" > /dev/null then iscsimount $mountpoint || overallstatus=$? fi done return $overallstatus } iscsiumountall() { # Find all fstab lines with iscsi as the fstype for mountpoint in `grep "iscsi," /etc/fstab | cut -f 2`; do # Only try to umount filesystems that are currently mounted. if mount | grep "on $mountpoint\b" > /dev/null then iscsiumount $mountpoint || overallstatus=$? fi done return $overallstatus } iscsicheckall() { # Find all fstab lines with iscsi as the fstype for mountpoint in `grep "iscsi," /etc/fstab | cut -f 2`; do iscsicheck $mountpoint || overallstatus=$? done return $overallstatus } case "$1" in start) $ISCSISCRIPT $@ && sleep 1 && iscsimountall ;; stop) iscsiumountall && $ISCSISCRIPT $@ ; overallstatus=$? ;; status) $ISCSISCRIPT $@ && iscsicheckall ;; restart|reload) $0 stop && sleep 1 && $0 start ; overallstatus=$? ;; initiator) shift ; $ISCSISCRIPT $@ ; overallstatus=$? ;; *) echo "Usage: $0 {start|stop|status|restart|reload|initiator}"; overallstatus=1 ;; esac rc_failed $overallstatus rc_exit August 14, 2009, at 12:52 PM
by -
Changed line 17 from:
->[+/etc/init.d/iscsi-mount+] to:
->[+/etc/fstab+] ''(an example)'' Added lines 23-29:
->[+/etc/init.d/iscsi-mount+] (:div style="border-style:ridge; border-width:2px; background-color:#ffffcc; margin-left:50px; overflow:auto; width:650px; height:300px;":) [@ @] (:divend:) August 14, 2009, at 12:52 PM
by -
Changed lines 5-12 from:
The problem is that you are unable to mount the iSCSI LUNs from within the [@/etc/fstab@] file. Now I am fully willing to admit that this "problem" is just due to my own stupidity, but from the research I have done online, I am not the only one to notice this problem. I also had some minor problems with permissions on the cert.b64 file (make sure radiusd user and group can see the file) If you run into any problems just run the following command to see all the necessary troubleshooting output: ->[@radiusd -A -X@] to:
The problem is that you are unable to mount the iSCSI LUNs from within the [@/etc/fstab@] file when using the open-scsi software initiator. Now I am fully willing to admit that this "problem" is just due to my own stupidity, but from the research I have done online, I am not the only one to notice this problem. The issue I had was that the system would try to mount the LUNs before the network was up and therefore failed. I tried a number of possible remedies but nothing worked: *Figuring out what remote_fs is and how to use it. *Moving the network_remote deamon script till after the iSCSI initiator was up. *Creating my own [@/etc/fstab.iscsi@] file. What other people have done is to create custom scripts. However my problem with these script is that they seemed sloppy. I needed a system that would handle both the iscsi initiator and the mounting of the filesystems in one go. And I also wanted a system that would work with the existing GNU utilities. What I can up with was a wrapper script for the open-scsi deamon (see below). This handles both the software initiator and mounting of the filesystems - for both starting and stopping. This way I can be sure certain that the system will neither try to mount filesystems before the an iSCSI connection is established nor kill the iSCSI connection without first umounting the filesystems. To make this script work it is necessary to slightly modify the iSCSI entries in the [@/etc/fstab@] file August 14, 2009, at 12:33 PM
by -
Changed line 13 from:
->[+/etc/raddb/radiusd.conf+] to:
->[+/etc/init.d/iscsi-mount+] August 14, 2009, at 12:26 PM
by -
Changed lines 3-11 from:
FreeRadius is a very versatile, powerful application for network RADIUS authentication. Perhaps too powerful! The problem is that is can seem to have too many options, and unless your a RADIUS expert, your screwed! What I needed for a customer was way to "proxy" LDAP/eDirectory authentication through RADIUS, and perhaps define redundant LDAP servers. That's all! Through a great deal of research, trial and error; I can up with the following slimmed down configuration file. Simple, Secure and Functional! Copy this file (with your specific modifications) and export your eDirectory Self-Signed CA (without private key) as Base64 to /etc/raddb/certs/cert.b64 You might also want to create a link to the init.d script. ->[@ln -s /etc/init.d/radiusd /usr/sbin/rcradiusd@] to:
I have been using iSCSI with SLES10 for a while without any problems. And even though I have been using SLES11 for months, I never ran across this problem until recently. Turns out that using iSCSI with SLES11 (or SuSE 11.1) has a bit of a problem. The problem is that you are unable to mount the iSCSI LUNs from within the [@/etc/fstab@] file. Now I am fully willing to admit that this "problem" is just due to my own stupidity, but from the research I have done online, I am not the only one to notice this problem. August 14, 2009, at 12:16 PM
by -
Changed lines 20-163 from:
############################################################################## # # # /etc/raddb/radiusd.conf # # # # Author: Bob Brandt (projects@brandt.ie) # # # # FreeRADIUS Redundant LDAP Configuration # # # ############################################################################## prefix = /usr exec_prefix = ${prefix} sysconfdir = /etc localstatedir = /var sbindir = ${exec_prefix}/sbin logdir = ${localstatedir}/log/radius raddbdir = ${sysconfdir}/raddb radacctdir = ${logdir}/radacct # Location of config and logfiles. confdir = ${raddbdir} run_dir = ${localstatedir}/run/radiusd log_file = ${logdir}/radius.log libdir = /usr/lib/freeradius pidfile = ${run_dir}/radiusd.pid user = radiusd group = radiusd max_request_time = 30 delete_blocked_requests = no cleanup_delay = 5 max_requests = 1024 bind_address = * port = 0 hostname_lookups = no allow_core_dumps = no regular_expressions = yes extended_expressions = yes log_stripped_names = yes log_auth = yes log_auth_badpass = yes log_auth_goodpass = yes usercollide = no lower_user = after lower_pass = no nospace_user = after nospace_pass = no checkrad = ${sbindir}/checkrad # SECURITY CONFIGURATION security { max_attributes = 200 reject_delay = 1 status_server = yes } # CLIENTS CONFIGURATION client 0.0.0.0/0 { secret = companysecret shortname = company } # THREAD POOL CONFIGURATION thread pool { start_servers = 5 max_servers = 32 min_spare_servers = 3 max_spare_servers = 10 max_requests_per_server = 0 } # LDAP Information ldap_identity = "cn=RadiusAdmin,ou=RemoteAccess,o=tree" ldap_password = thisisnottherealpassword ldap_basedn = "o=tree" ldap_tls_cacertfile = "/etc/raddb/certs/cert.b64" # MODULE CONFIGURATION modules { ldap ldap1 { server = "ldap1.domain.name" identity = ${ldap_identity} password = ${ldap_password} basedn = ${ldap_basedn} filter = "(&(objectClass=inetOrgPerson)(cn=%{Stripped-User-Name:-%{User-Name}}))" base_filter = "(objectclass=inetOrgPerson)" start_tls = yes tls_cacertfile = ${ldap_tls_cacertfile} tls_require_cert = "demand" access_attr = "dialupAccess" dictionary_mapping = ${raddbdir}/ldap.attrmap ldap_connections_number = 5 password_attribute = nspmPassword timeout = 4 timelimit = 3 net_timeout = 1 edir_account_policy_check = yes } ldap ldap2 { server = "ldap2.domain.name" identity = ${ldap_identity} password = ${ldap_password} basedn = ${ldap_basedn} filter = "(&(objectClass=inetOrgPerson)(cn=%{Stripped-User-Name:-%{User-Name}}))" base_filter = "(objectclass=inetOrgPerson)" start_tls = yes tls_cacertfile = ${ldap_tls_cacertfile} tls_require_cert = "demand" access_attr = "dialupAccess" dictionary_mapping = ${raddbdir}/ldap.attrmap ldap_connections_number = 5 password_attribute = nspmPassword timeout = 4 timelimit = 3 net_timeout = 1 edir_account_policy_check = yes } } authorize { ldap1 ldap2 } authenticate { Auth-Type LDAP { ldap1 ldap2 } } post-auth { Post-Auth-Type REJECT { ldap1 ldap2 } } to:
Deleted lines 22-29:
References: * %newwin% http://www.novell.com/coolsolutions/tip/15922.html * %newwin% http://www.novell.com/documentation/edir_radius/index.html * %newwin% http://en.opensuse.org/RadiusServerHOWTO August 14, 2009, at 12:15 PM
by -
Added lines 1-173:
(:title Workaround for iSCSI issues on SLES11:) FreeRadius is a very versatile, powerful application for network RADIUS authentication. Perhaps too powerful! The problem is that is can seem to have too many options, and unless your a RADIUS expert, your screwed! What I needed for a customer was way to "proxy" LDAP/eDirectory authentication through RADIUS, and perhaps define redundant LDAP servers. That's all! Through a great deal of research, trial and error; I can up with the following slimmed down configuration file. Simple, Secure and Functional! Copy this file (with your specific modifications) and export your eDirectory Self-Signed CA (without private key) as Base64 to /etc/raddb/certs/cert.b64 You might also want to create a link to the init.d script. ->[@ln -s /etc/init.d/radiusd /usr/sbin/rcradiusd@] I also had some minor problems with permissions on the cert.b64 file (make sure radiusd user and group can see the file) If you run into any problems just run the following command to see all the necessary troubleshooting output: ->[@radiusd -A -X@] ->[+/etc/raddb/radiusd.conf+] (:div style="border-style:ridge; border-width:2px; background-color:#ffffcc; margin-left:50px; overflow:auto; width:650px; height:300px;":) [@ ############################################################################## # # # /etc/raddb/radiusd.conf # # # # Author: Bob Brandt (projects@brandt.ie) # # # # FreeRADIUS Redundant LDAP Configuration # # # ############################################################################## prefix = /usr exec_prefix = ${prefix} sysconfdir = /etc localstatedir = /var sbindir = ${exec_prefix}/sbin logdir = ${localstatedir}/log/radius raddbdir = ${sysconfdir}/raddb radacctdir = ${logdir}/radacct # Location of config and logfiles. confdir = ${raddbdir} run_dir = ${localstatedir}/run/radiusd log_file = ${logdir}/radius.log libdir = /usr/lib/freeradius pidfile = ${run_dir}/radiusd.pid user = radiusd group = radiusd max_request_time = 30 delete_blocked_requests = no cleanup_delay = 5 max_requests = 1024 bind_address = * port = 0 hostname_lookups = no allow_core_dumps = no regular_expressions = yes extended_expressions = yes log_stripped_names = yes log_auth = yes log_auth_badpass = yes log_auth_goodpass = yes usercollide = no lower_user = after lower_pass = no nospace_user = after nospace_pass = no checkrad = ${sbindir}/checkrad # SECURITY CONFIGURATION security { max_attributes = 200 reject_delay = 1 status_server = yes } # CLIENTS CONFIGURATION client 0.0.0.0/0 { secret = companysecret shortname = company } # THREAD POOL CONFIGURATION thread pool { start_servers = 5 max_servers = 32 min_spare_servers = 3 max_spare_servers = 10 max_requests_per_server = 0 } # LDAP Information ldap_identity = "cn=RadiusAdmin,ou=RemoteAccess,o=tree" ldap_password = thisisnottherealpassword ldap_basedn = "o=tree" ldap_tls_cacertfile = "/etc/raddb/certs/cert.b64" # MODULE CONFIGURATION modules { ldap ldap1 { server = "ldap1.domain.name" identity = ${ldap_identity} password = ${ldap_password} basedn = ${ldap_basedn} filter = "(&(objectClass=inetOrgPerson)(cn=%{Stripped-User-Name:-%{User-Name}}))" base_filter = "(objectclass=inetOrgPerson)" start_tls = yes tls_cacertfile = ${ldap_tls_cacertfile} tls_require_cert = "demand" access_attr = "dialupAccess" dictionary_mapping = ${raddbdir}/ldap.attrmap ldap_connections_number = 5 password_attribute = nspmPassword timeout = 4 timelimit = 3 net_timeout = 1 edir_account_policy_check = yes } ldap ldap2 { server = "ldap2.domain.name" identity = ${ldap_identity} password = ${ldap_password} basedn = ${ldap_basedn} filter = "(&(objectClass=inetOrgPerson)(cn=%{Stripped-User-Name:-%{User-Name}}))" base_filter = "(objectclass=inetOrgPerson)" start_tls = yes tls_cacertfile = ${ldap_tls_cacertfile} tls_require_cert = "demand" access_attr = "dialupAccess" dictionary_mapping = ${raddbdir}/ldap.attrmap ldap_connections_number = 5 password_attribute = nspmPassword timeout = 4 timelimit = 3 net_timeout = 1 edir_account_policy_check = yes } } authorize { ldap1 ldap2 } authenticate { Auth-Type LDAP { ldap1 ldap2 } } post-auth { Post-Auth-Type REJECT { ldap1 ldap2 } } @] (:divend:) References: * %newwin% http://www.novell.com/coolsolutions/tip/15922.html * %newwin% http://www.novell.com/documentation/edir_radius/index.html * %newwin% http://en.opensuse.org/RadiusServerHOWTO |