![]() Get_IORegistry_info: find_IORegistryBSDName for returned 2įilesystem NOT mounted on /Volumes/volname ![]() I got a strange output pasted below which seems to indicate there is a potential issue with fstab, which may also explain the auto mount on boot not working:Ġ : = IOMedia I then unmounted the volume, created the /Volumes/volname directory manually and issued "mount -a" in order to have the mount routine use the options listed in fstab - in this case, only the mount path being specified, all else just standard (rw and acfs). The command successfully mounted the volume, and again automatically altered the fstab to include the specified option, in this case mount point, and removed all other options. I took my fstab experiment one step further and this is where I hit a wall. I believe the options would only show in mount output and have an effect if the volume was mounted via fstab (as it would during boot), so.Ĥ. I then rebooted as I did during every other test, never did the volume mount on boot, and no increase in log verbosity. I also watched the system and nssdbg logs, and did not notice any uptick in verbosity or debug info. Next i used the "mount" command to see if the options would be displayed for this mount, but they weren't, only "acfs" and "local" were there. LABEL=volname none acfs rw,mnt_retrans=10,mnt_recon=hard,mnt_type=bg,verbose=yes,debug=yes Which mounted the volume successfully and automatically altered the fstab line to look like this: Xsanctl mount volname -mnt_retrans=10 -mnt_recon=hard -mnt_type=bg -verbose=yes -debug=yes" ![]() I went a step further and used xsanctl to add several options to fstab (in order to see if I could get more info and potentially make the client continue attempting to mount once the LUNs are found) but I could never get those options to be respected or show an effect. These 3 LUNs should be able to be scanned plenty fast enough to register the LUNs before the mount attempt unless there is an order of operations issue in the code.ģ. As a test we zoned out all other storage controllers except the one chassis that contained the MDJ, and 2 LUNs that make up the one volume. After some troubleshooting, I now doubt this is the issue, or the *only* issue I should say. I thought that because in the nssdbg.out the disk scan successfully finding the LUNs always happens just after the "new mount registered" entry (sometimes within the same second, which may mean its just a display output that is run after the mount attempt happens). At first I thought maybe the LUNs weren't being scanned by the time the mount attempt happens, triggering the failure (a long standing issue with macs, except they usually retry until successful), and that there may be something needed in order to get it to retry after all LUNs are found. Looking at the nssdbg log after booting and the volume not mounting shows nothing obviously relevant. This indicates there is nothing wrong with the profile, the volume, zoning, etc and the auto mount feature "works" if the system is already booted and I load the profile at that point.Ģ. When I install the profile, the expected volume mounts immediately, no others do, and everything is perfect. There is a profile loaded with both xsan and preference payloads, and onlyMount key set to auto mount one volume (out of several). The problem is only getting the volume to mount on boot, everything else is working fine. ![]() I've been having trouble with auto mount on a 10.11.2 client mac client (the only El Capitan client I've tested with Xsan MDCs) - and dug fairly deep, unsuccessfully.
0 Comments
Leave a Reply. |
AuthorWrite something about yourself. No need to be fancy, just an overview. ArchivesCategories |