Red Hat Bugzilla – Bug 602948
SRP LUNs not found at boot time
Last modified: 2010-08-02 10:20:21 EDT
Description of problem:
In RHEL 6 beta 1, SRP LUNs that should be found and attached to the server at boot time aren't.
Instead, the user has to manually run a script or create their own boot script to achieve this.
Version-Release number of selected component (if applicable):
Steps to Reproduce:
1. Configure remote SRP targets for access by host server.
2. Verify connectivity from host server using "ibsrpdm -c" and/or "srp_daemon -eo".
3. Add LUN wwid's to host device mapper multipath configuration file /etc/multipath.conf file
4. Enable SRP module in /etc/rdma/rdma.conf on host.
5. Use chkconfig on host to enable rdma service at boot.
6. Reboot. The LUNs are not found. (This is the problem)
No SRP LUNs available on the host. Subsequent scripts and programs depending upon them then fail.
SRP LUNs to be attached to the host and made available, as per other storage types. (ie iSCSI, FC, etc)
Do you reckon this could be resolved before RHEL 6 final ships?
SRP configuration information is presently being written for inclusion in the RHEL 6 Virtualization Guide. Thinking it would be better to have this "just work" rather than instruct end users to create their own init script to manage this.
This request was evaluated by Red Hat Product Management for inclusion in a Red
Hat Enterprise Linux major release. Product Management has requested further
review of this request by Red Hat Engineering, for potential inclusion in a Red
Hat Enterprise Linux Major release. This request is not yet committed for
Created attachment 423177 [details]
Patch against the rdma init script
Attached is a patch against the latest rdma init.d script I could find (in Fedora 14 rawhide).
The first addition to start() is pretty much lifted from the OFED 1.5.2-rc1 version, introducing a new "SRP_DAEMON_ENABLE=yes" variable in your /etc/rdma/rdma.conf.
The OFED code for flushing the multipath device maps in stop() was completely non-functional however, so that's been re-worked. It now correctly retrieves device map names and flushes them, allowing the ib_srp module to be nicely removed.
This is tested and working on the local SRP connected lab systems here.
This issue has been proposed when we are only considering blocker
issues in the current Red Hat Enterprise Linux release. It has
been denied for the current Red Hat Enterprise Linux release.
** If you would still like this issue considered for the current
release, ask your support representative to file as a blocker on
your behalf. Otherwise ask that it be considered for the next
Red Hat Enterprise Linux release. **
*** This bug has been marked as a duplicate of bug 591169 ***