Comments (8)
  • Add a Comment
  • Edit
  • More Actions v
  • Quarantine this Entry

1 Mach_v commented Permalink

(i'm tech at ibm BP) it's strange, but after read this notes, i'm contact with one of our customers who own v7000, he purchase it year ago, and update to after it available. So Test Utility tool now say - uptime 248!!! day at node1 and node2. <div>&nbsp;</div> i'm open ticket at ibm support, and will wait for more info about this. <div>&nbsp;</div> i'm wonder, do i'm need push customer to upgrade, or wait for more info about it.

2 anthonyv commented Permalink

Wait for feedback from IBM Support... I would like to know if there is something different about that level, a problem with the tool or if the customer if just plain lucky.

3 Mach_v commented Permalink

official answer of hwsupport - upgrade to <br /> About 248 day of uptime - answer there is *possible* restart after 208 day. <div>&nbsp;</div> I'm not happy with this level of support - possible restart instead of *will restart* as noted at alert, no 6.3 update to avoid 2 stage upgrade ( -&gt; -&gt; 6.3.0.X). <div>&nbsp;</div> Customer too not happy with wrong alarm, and thinks about avoid update and wait for 6.3. <div>&nbsp;</div> If first alarm say about possible failure after 208 day there been no misunderstandings.

4 anthonyv commented Permalink

Its a very good point: The difference between 'will' and 'may'. <div>&nbsp;</div> I checked the Linux posts like these: <br />;sliceId=1 <br /> Which says: "Possible system freeze after more than 208 days" It does not say 'definite' freeze. <div>&nbsp;</div> and this one: <br /> <div>&nbsp;</div> I am not sure which one applies to IBMs Linux kernel, but the point being that in each case there is the risk of lock up and crash. <br /> I suppose in fairness to IBM, the alert is to prevent a possible outage so the language needs to be strong to encourage the clients to take action. I hate to use this word.... but your client has been lucky.... and we should not rely on luck. They should still take action on this issue. <br />

5 Mach_v commented Permalink

I'm read linux kernel bug description, and wonder do i'll able to simple upgrade to without first reboot controller, people in bug list say about hang when try "apt upgrade" after 208 day mark. Do i'll need first reboot both controller (with hour interval between) and then upgrade, or try upgrade without reboot? i'm not sure customer will follow ( if it try upgrade without reboot first at friday evening (maintenance window with minimal i/o) while he have only 8x5xBD software support :(

6 anthonyv commented Permalink

In your case I would definitely NOT do the upgrade till you have done a rolling reboot of the nodes. Once both nodes report 0 days uptime.... then do the firmware upgrade,

7 Mach_v commented Permalink

thanks for clearance of this issue, hwsupport fail even notice this when close ticket :(

8 anthonyv commented Permalink

The iBM Web Alert actually got updated to reflect my advice/. Just because you got past 208 days does not mean you are totally in the clear. You will need to do a rolling reboot of each node.