<html xmlns:v="urn:schemas-microsoft-com:vml" xmlns:o="urn:schemas-microsoft-com:office:office" xmlns:w="urn:schemas-microsoft-com:office:word" xmlns:st1="urn:schemas-microsoft-com:office:smarttags" xmlns="http://www.w3.org/TR/REC-html40"> <head> <meta http-equiv=Content-Type content="text/html; charset=us-ascii"> <meta name=Generator content="Microsoft Word 11 (filtered medium)"> <o:SmartTagType namespaceuri="urn:schemas-microsoft-com:office:smarttags" name="PersonName"/>  <style>  </style> </head> <body lang=EN-US link=blue vlink=purple> <div class=Section1> In the previous logs, the multipath daemon wasn’t running. Attached is the new log file, which has the path checker logs too.<o:p></o:p> -Murthy<o:p></o:p> <o:p> </o:p> <div> <div class=MsoNormal align=center style='text-align:center'> <hr size=2 width="100%" align=center tabindex=-1> </div> From: <st1:PersonName w:st="on">Murthy, Narasimha Doraswamy</st1:PersonName> (STSD) Sent: Friday, June 17, 2005 1:02 PM To: '<st1:PersonName w:st="on">dm-devel@redhat.com</st1:PersonName>' Subject: regarding path failback using HSV110 array.<o:p></o:p> </div> <o:p> </o:p> Hi Alasdair,<o:p></o:p> As per yesterday’s teleconf, I am sending the snapshot of the /var/log/messages, which has the messages logged when testing the basic failback functionality. The testing is conducted on the host with RHEL U1 and single controller of HP Storageworks HSV110 array. As each HSV110 controller has 2 ports, we effectively see 2 paths per lun. I am using 0.4.5.44 version of multipath tools. <o:p></o:p> <o:p> </o:p> Following is the observation during testing the path failback.<o:p></o:p> <o:p> </o:p> a) When the path is made unavailable (/dev/sdd) for a logical device “mp1” on a multipath target, the status of the path becomes [faulty][active]<o:p></o:p> <o:p> </o:p> mp1<o:p></o:p> [size=1 GB][features="0"][hwhandler="0"]<o:p></o:p> \_ round-robin 0 [active][best]<o:p></o:p> \_ 0:0:0:1 sda 8:0 [faulty][active]<o:p></o:p> \_ round-robin 0 [enabled]<o:p></o:p> \_ 1:0:0:1 sdd 8:48 [ready][active]<o:p></o:p> <o:p> </o:p> b) When an dd is executed on /dev/mapper/mp1 (dd if=/dev/zero of=/dev/devmapper/mp1), the state of the paths is as follows<o:p></o:p> mp1<o:p></o:p> [size=1 GB][features="0"][hwhandler="0"]<o:p></o:p> \_ round-robin 0 [enabled]<o:p></o:p> \_ 0:0:0:1 sda 8:0 [faulty][failed]<o:p></o:p> \_ round-robin 0 [active][best]<o:p></o:p> \_ 1:0:0:1 sdd 8:48 [ready ][active]<o:p></o:p> <o:p> </o:p> <o:p> </o:p> c) Now if the path (/dev/sda) is made available for the same logical device, the status is as follows<o:p></o:p> mp1<o:p></o:p> [size=1 GB][features="0"][hwhandler="0"]<o:p></o:p> \_ round-robin 0 [enabled]<o:p></o:p> \_ 0:0:0:1 sda 8:0 [ready ][failed]<o:p></o:p> \_ round-robin 0 [active][best]<o:p></o:p> \_ 1:0:0:1 sdd 8:48 [ready ][active]<o:p></o:p> <o:p> </o:p> Note the path “/dev/sda” has become [ready][failed] not [ready][active]. For this to change to [ready][active], “multipath –v0” needs to be executed manually.<o:p></o:p> <o:p> </o:p> d) Now if path “/dev/sdd” is made unavailable, then if “dd” is run on /dev/mapper/mp1, it returns with input/output error, as none of the path will be in [ready][active] state. The following is the state now:<o:p></o:p> <o:p> </o:p> mp1<o:p></o:p> [size=1 GB][features="0"][hwhandler="0"]<o:p></o:p> \_ round-robin 0 [enabled][best]<o:p></o:p> \_ 0:0:0:1 sda 8:0 [ready ][failed]<o:p></o:p> \_ round-robin 0 [enabled]<o:p></o:p> \_ 1:0:0:1 sdd 8:48 [faulty][failed]<o:p></o:p> <o:p> </o:p> <o:p> </o:p> If “multipath –v0” executed from the command prompt the states get updated properly as<o:p></o:p> <o:p> </o:p> mp1<o:p></o:p> [size=1 GB][features="0"][hwhandler="0"]<o:p></o:p> \_ round-robin 0 [active][best]<o:p></o:p> \_ 0:0:0:1 sda 8:0 [ready ][active]<o:p></o:p> \_ round-robin 0 [enabled]<o:p></o:p> \_ 1:0:0:1 sdd 8:48 [faulty][active]<o:p></o:p> <o:p> </o:p> After this I will be able to do IO using “dd”. How do we make the path state update properly with out manually invoking “multipath –v0”? <o:p></o:p> I am not sure on how to get the debug build of multipath and multipathd binaries, if the information in the messages.txt is insufficient, I need to test with the debug build and send the new log.<o:p></o:p> <o:p> </o:p> Any help in resolving this problem is appreciated.<o:p></o:p> <o:p> </o:p> Thanks and regards<o:p></o:p> Murthy<o:p></o:p> <o:p> </o:p> </div> </body> </html>