I have looked at your logs.. Here are the messages.. Mar 3 15:38:37 rman-ss-03 kernel: sd 1:0:0:0: Done: SUCCESS Mar 3 15:38:37 rman-ss-03 kernel: 2 sd 1:0:0:0: Mar 3 15:38:37 rman-ss-03 kernel: command: Synchronize Cache(10): 35 00 00 00 00 00 00 00 00 00 Mar 3 15:38:37 rman-ss-03 kernel: : Current: sense key: Illegal Request Mar 3 15:38:37 rman-ss-03 kernel: <<vendor>> ASC=0x94 ASCQ=0x1ASC=0x94 ASCQ=0x1 Mar 3 15:38:37 rman-ss-03 kernel: Mar 3 15:38:38 rman-ss-03 kernel: sd 1:0:0:0: Done: SUCCESS Mar 3 15:38:38 rman-ss-03 kernel: 2 sd 1:0:0:0: Mar 3 15:38:38 rman-ss-03 kernel: command: Synchronize Cache(10): 35 00 00 00 00 00 00 00 00 00 Mar 3 15:38:38 rman-ss-03 kernel: : Current: sense key: Illegal Request Mar 3 15:38:38 rman-ss-03 kernel: <<vendor>> ASC=0x94 ASCQ=0x1ASC=0x94 ASCQ=0x1 These messages are coming because of Synchronize Cache commands. It may be related to type of filesystem you are on. Did you have barrier enabled on your filesystem. I see that these commands are going out to target on passive paths.. I dont understand that part yet. Normally scsi device handlers block commands going out passive paths..This should be coming in a differrent path.. I still have to figure that out.. ________________________________________ From: dm-devel-bounces@xxxxxxxxxx [dm-devel-bounces@xxxxxxxxxx] On Behalf Of Charles Slivkoff [slivkoff@xxxxxxx] Sent: Thursday, March 03, 2011 2:59 PM To: device-mapper development Subject: Re: Dell MD3200 - sd: Current: sense key: Illegal Request ASC=0x94 ASCQ=0x1ASC=0x94 ASCQ=0x1 On 03/03/2011 11:47 AM, Menny_Hamburger@xxxxxxxx wrote: > > I know this is usually a wild goose chase, but what about the > controller firmware? I can see from the profile that the controller > FW is 07.70.06.63, while the latest according to the Dell > compatibility matrix is 07.75.14.60 I'll look into this next. On 03/03/2011 01:30 PM, Yanqing_Liu@xxxxxxxx wrote: > > It seems that you mapped your LUNs some time after host bootup. That > causes pseudo LUN 0 to be created. If you lay down DM using Dell > software package, please invoke "rescan_dm_devs" for > re-configuration. I have used "rescan_dm_devs" before. This did not eliminate the pseudo LUN 0. I did manage to get rid of them, though, by un-mapping and re-mapping the array virtual disks as LUNs 0 & 1, instead of LUNs 1 & 2. The current array profile is here: http://tinypaste.com/64eebb lsscsi now looks good. See http://tinypaste.com/3580a Unfortunately, this did not help with the "Illegal Request" messages. > We may need your related kernel message for further analysis. As to > why setting "failback" to manual doesn't work on MD3200, the problem > may be in multipathing daemon, in combination with SAS tranpportation > trearing down all devices immediately after path loss. We've filed a > bugzilla for this. The latest syslog is here: http://tinypaste.com/87516 > BTW, what is your LVM setting for these devices? The LUNs are each assigned to their own SCSI controller. Both LUNs (PVs) are members of a single VG. The LVs are striped across the PVs, thus allowing both paths to be used. The "lvdisplay -m" output is here: http://tinypaste.com/7c42a -- dm-devel mailing list dm-devel@xxxxxxxxxx https://www.redhat.com/mailman/listinfo/dm-devel -- dm-devel mailing list dm-devel@xxxxxxxxxx https://www.redhat.com/mailman/listinfo/dm-devel
