Start a Conversation
UNSOLVED

sk_

updated

13 years ago

S

sk_

1 Rookie

•

104 Posts

0

2537

January 23rd, 2014 06:00

VNX5400 and ESX multipath issues?

Hi,

using VNX5400 block only, FC san and ESX5.x clusters with native multipath. Policies are default, satp_alua_cx / RoundRobin.

We have noticed that when doing SP reboots or disabling one HBA link on host/switch causes "Lost Acces to volume" errors on ESX and it will take 30-60 seconds to reconnect.This did not happen with our older CX4 systems.

Anyone seen similar probs with new VNX2?

sk

  • S

    srumpel

    15 Posts

    761

    0

    Posted January 24th, 2014 14:00

    Hi,

    do you run the latest MCX code 05.33.000.5.038?

  • S

    sk_

    1 Rookie

    •

    104 Posts

    761

    0

    Posted January 25th, 2014 01:00

    nope, .035 at the moment and original release before that.

    .038 upgrade next week, but didn't notice anything relating to this in the docs?

    sk

  • S

    srumpel

    15 Posts

    761

    0

    Posted January 25th, 2014 17:00

    what FC switch vendor is used and was your CX4 system 4G or 8G version before?

    I using VNX5200 with vSphere 5.5 without such issue.

  • S

    sk_

    1 Rookie

    •

    104 Posts

    761

    0

    Posted January 26th, 2014 01:00

    now cisco 8G, earlier cisco 4G.

    When you do SP reboot on 5200, what events do you see on the ESX servers tasks&events list?

    Do you get "Lost access to volume xx" and "Restored access to volume xx" after that, and what is the time between these?

    I assume u are using native multipath and not PowerPath? (and with defaults, satp_alua_cx / RoundRobin?)

    tnx,

    sk

  • S

    srumpel

    15 Posts

    761

    0

    Posted January 26th, 2014 05:00

    Hello Sami,

    I will do SP reboot Wednesday and look whats happen.

    I using native multipath with satp_alua_cx / RoundRobin like you

    Do you use Failover Mode 4 on VNX for host initiator?

  • S

    sk_

    1 Rookie

    •

    104 Posts

    761

    0

    Posted January 27th, 2014 03:00

    Yes, failovermode4 on ESX5.x hosts.

    Just noticed that ESX4.1 hosts with failovermode1 / MRU doesn't have this problem, they work ok..

    (did the 0.38 upgrade today)

  • S

    srumpel

    15 Posts

    761

    0

    Posted January 29th, 2014 11:00

    Hello Sami,

    today reboot both SP several time and can't reproduce your issue.

    no volume disconnect accour and vxpa.log looks good.

  • S

    sk_

    1 Rookie

    •

    104 Posts

    761

    0

    Posted January 29th, 2014 22:00

    Ok, thanks for reporting!

    I assume you then got just the "Path redundancy lost" alerts when SP went down?

    btw. I have looked these from VIclient - task&events - event list.. logfiles shows a bit different errors.

Dell Support Resources

Latest Solutions

View More