HPE Storage Users Group https://3parug.com/ |
|
Failover behaviour vs. ESXi APDs https://3parug.com/viewtopic.php?f=18&t=2586 |
Page 2 of 2 |
Author: | mlu [ Tue Aug 08, 2017 4:07 am ] |
Post subject: | Re: Failover behaviour vs. ESXi APDs |
koopa79 wrote: Have you tried logging a call with vmware as well? Yes, VMware says everything OK from their side. ESX keeps trying to connect through the paths of the primary storage, but receives "Target Busy" states, eventually they come back as ok in the end. According to VMW engineer "busy state" (see SCSI sense codes 2nd post from "Wed Jul 19, 2017 11:18" further up this thread) can be caused not only by the target (storage) itself, but because of the HBA too. Currently it looks as if it works fine with qLogic HBAs but fails with Emulex HBAs. FW & drivers are in accordance with the VMware and HPE certified versions. Thanks for efforts so far though. |
Author: | ailean [ Tue Aug 08, 2017 7:22 am ] |
Post subject: | Re: Failover behaviour vs. ESXi APDs |
mlu wrote: koopa79 wrote: Have you tried logging a call with vmware as well? Yes, VMware says everything OK from their side. ESX keeps trying to connect through the paths of the primary storage, but receives "Target Busy" states, eventually they come back as ok in the end. According to VMW engineer "busy state" (see SCSI sense codes 2nd post from "Wed Jul 19, 2017 11:18" further up this thread) can be caused not only by the target (storage) itself, but because of the HBA too. Currently it looks as if it works fine with qLogic HBAs but fails with Emulex HBAs. FW & drivers are in accordance with the VMware and HPE certified versions. Thanks for efforts so far though. Cheers for the update, we've run similar tests when new arrays went in and didn't see any issues however we've used qlogic HBAs exclusively for many years. It's not the first time I've seen people having odd issues that relate to Emulex, hopefully HPE/VMware can poke Emulex for a driver/fw update for this. |
Author: | mlu [ Fri Aug 11, 2017 6:35 am ] |
Post subject: | Re: Failover behaviour vs. ESXi APDs |
Still unclear. We went that far now that it seems as HPE is going to rebuild the environment in their lab. Stay tuned. |
Author: | drpixel76 [ Mon Oct 09, 2017 4:07 am ] |
Post subject: | Re: Failover behaviour vs. ESXi APDs |
Hello, Do you have any update ? I can confirm the different behavior between Qlogic and Emulex. When I do a MTF (Manual Transparent Failover) it works great but with ATP (Automatic Transparent Failover) I get a APD. The same ESXi with a Qlogic HBA is working with MTF and ATF. So it seems to have an (undocumented !?) issue with Emulex/3PAR Peer Peristence /VMware PS : I can get the test working with Emulex by also disabling the SAN switch host port connected to the 3PAR that is "broken". Best Regards, |
Author: | bajorgensen [ Mon Oct 09, 2017 6:39 am ] |
Post subject: | Re: Failover behaviour vs. ESXi APDs |
Have you disabled persistent checksum on the Emulex card as per the SPOCK? (if it is 16G or 32G cards) - Anders |
Author: | drpixel76 [ Tue Oct 10, 2017 5:09 am ] |
Post subject: | Re: Failover behaviour vs. ESXi APDs |
Hello, Yes it has been done. But the issue is still present. Best Regards, |
Page 2 of 2 | All times are UTC - 5 hours |
Powered by phpBB © 2000, 2002, 2005, 2007 phpBB Group http://www.phpbb.com/ |