https://bugzilla.novell.com/show_bug.cgi?id=233546 Summary: IO discontinued to some of the LUNs after one of the host / filer side cable pulls Product: SUSE Linux 10.1 Version: Final Platform: i686 OS/Version: SLES 10 Status: NEW Severity: Blocker Priority: P5 - None Component: Basesystem AssignedTo: bnc-team-screening@forge.provo.novell.com ReportedBy: rsarraf@netapp.com QAContact: qa@suse.de CC: xdl-novell-bugzilla@netapp.com * Discover 5 luns from the filer to the host. * Since filer is having 4 network interfaces, each lun will have 4 paths. * Create filesystem ( reiserfs or ext3) on all 5 multipath devices. * Run IO on all 5 devices. * Remove or down any of the 4 interfaces. * IO stalled on all the devices for around 180 sec. * IO didn't resume on all the luns. * IO process goes into uninterruptible sleep mode, which cannot be killed Upon removal on one of the cables, the IO gets on hold for a while (180 seconds). Upon resume, after the threshold, the IO doesn't continue on all of the LUNs. Most of the LUNs don't receive IO at all. Another odd behavior is the output from multipath -l, which reports all the paths as active for the LUNs which weren't receiving IO. Putting it as an example, when I removed the cables from one of the ports, IO got discontinued on LUNs 2,3 and 4. IO was still continuing on LUNs 1 and 5 (We're using iostat to see the statistics for IO). But when I used multipath -l, it reported that: * One path was failed for the LUNs which were receiving IO i.e. LUNs 1 and 5 * For the remaining LUNs i.e. 2, 3 and 4 (which weren't receiving IO after the cable pull), multipath -l reported that all the paths were active. This is odd behavior. This behavior also doesn't change if the cable is pulled back in. -- Configure bugmail: https://bugzilla.novell.com/userprefs.cgi?tab=email ------- You are receiving this mail because: ------- You are on the CC list for the bug, or are watching someone who is.