Has anyone experienced an issue going from 7.0.6 to 7.2.5.1. I did an upgrade on my 940 cluster and when I attempted takeover the takeover never succeeded. The first head decided to reboot (without me running halt) but the second head never completed the takeover.
After that it complained about version mismatch.
Hi, Jerry...
It sounds like you attempted a nondisruptive (or rolling) upgrade, which is supported from 7.0.6 to 7.2.5.1 for most configurations. If you could provide additional detail on your system configuration as well as the steps you performed, we'll be better able to diagnose the issue.
Take care,
Steve Lawler Technical Marketing Engineer High Availability
NetApp 408.822.1486 steve.lawler@netapp.com www.netapp.com
-----Original Message----- From: Jerry [mailto:juanino@yahoo.com] Sent: Friday, 08 August, 2008 15:50 To: list toasters Subject: 7.0.6 to 7.2.5.1 upgrade
Has anyone experienced an issue going from 7.0.6 to 7.2.5.1. I did an upgrade on my 940 cluster and when I attempted takeover the takeover never succeeded. The first head decided to reboot (without me running halt) but the second head never completed the takeover.
After that it complained about version mismatch.
Hi Jerry,
Are the heads at exactly the same software revision ?
Do they run normally when you're not attempting a cluster failover ?
Are the partner configs match for the interfaces ?
There is a cluster failover test script you can run which will test all the pre-requisites for successful failover - I think you can download it from NOW.
Cheers, Raj.
That sometimes happens if you are attempting a non-disruptive upgrade and either miss a step, or something (download for example) does not complete as it should.
So, if you decide not to "cf disable", install, download, and reboot both heads individually (typical DISRUPTIVE upgrade), you have to follow steps simliar to the ones below.
1. Install software to both nodes 2. "download" on both nodes 3. Run "version -b" to ensure that both nodes have the new kernel version on the primary cf partition 4. filer1> cf takeover (filer2 now reboots, and you should notice the new kernel version mentioned. It then waits for giveback) 5. filer1> cf giveback -f (filer2 continues to boot) 6. At this point, filer2 is on the new ontap version, filer1 is downrev (takeover not possible) 7. filer1> halt (then immediately enter the following command on filer2) 8. filer2> cf takeover -n (-n switch used to allow NDU upgrade) 9. allow filer1 to boot and wait for giveback.. Note it should be on the new Ontap kernel at this point. 10. filer2> cf giveback -f 11. Check active version on each filer filer*> version
Note that these are the "basic" steps to perform a non-disruptive upgrade. More of an outline really. You should check the NOW site to ensure that Ontap versions involved are NDU capable, as well as any potential issues if SAN is involved.
Regards,
Michael Strickland Netapp Support
-----Original Message----- From: Jerry [mailto:juanino@yahoo.com] Sent: Friday, August 08, 2008 6:50 PM To: list toasters Subject: 7.0.6 to 7.2.5.1 upgrade
Has anyone experienced an issue going from 7.0.6 to 7.2.5.1. I did an upgrade on my 940 cluster and when I attempted takeover the takeover never succeeded. The first head decided to reboot (without me running halt) but the second head never completed the takeover.
After that it complained about version mismatch.
I used the upgrade advisor plans sent by netapp. What is interesting is I never remember in past upgrades doing download on both filers. (is this something new?). I basically did this:
filer1> download (on one filer) filer1> version -b filer2> cf takeover
--- note, filer1 immediately rebooted here, with no option for me to issue a halt ----
as filer1 was coming back up, I noticed filer2 never completed his takeover. once filer1 was back up cf status reported cf disabled because of version mismatch.
So, was filer1 supposed to reboot right away when filer2 did a takeover? I am trying to figure out what I did wrong here, but I think I did all the steps ok. On the update of filer2 I did not have a problem, as I did a cf takeover -n there which allowed me to run halt there.
To answer the other person's post, failover has worked about a week prior so it's a known working cluster.
--- On Sat, 8/9/08, Strickland, Michael Michael.Strickland@netapp.com wrote:
From: Strickland, Michael Michael.Strickland@netapp.com Subject: RE: 7.0.6 to 7.2.5.1 upgrade To: juanino@yahoo.com, "list toasters" toasters@mathworks.com Date: Saturday, August 9, 2008, 2:01 PM That sometimes happens if you are attempting a non-disruptive upgrade and either miss a step, or something (download for example) does not complete as it should.
So, if you decide not to "cf disable", install, download, and reboot both heads individually (typical DISRUPTIVE upgrade), you have to follow steps simliar to the ones below.
- Install software to both nodes
- "download" on both nodes
- Run "version -b" to ensure that both nodes
have the new kernel version on the primary cf partition 4. filer1> cf takeover (filer2 now reboots, and you should notice the new kernel version mentioned. It then waits for giveback) 5. filer1> cf giveback -f (filer2 continues to boot) 6. At this point, filer2 is on the new ontap version, filer1 is downrev (takeover not possible) 7. filer1> halt (then immediately enter the following command on filer2) 8. filer2> cf takeover -n (-n switch used to allow NDU upgrade) 9. allow filer1 to boot and wait for giveback.. Note it should be on the new Ontap kernel at this point. 10. filer2> cf giveback -f 11. Check active version on each filer filer*> version
Note that these are the "basic" steps to perform a non-disruptive upgrade. More of an outline really. You should check the NOW site to ensure that Ontap versions involved are NDU capable, as well as any potential issues if SAN is involved.
Regards,
Michael Strickland Netapp Support
-----Original Message----- From: Jerry [mailto:juanino@yahoo.com] Sent: Friday, August 08, 2008 6:50 PM To: list toasters Subject: 7.0.6 to 7.2.5.1 upgrade
Has anyone experienced an issue going from 7.0.6 to 7.2.5.1. I did an upgrade on my 940 cluster and when I attempted takeover the takeover never succeeded. The first head decided to reboot (without me running halt) but the second head never completed the takeover.
After that it complained about version mismatch.
Hi, Jerry...
Yes, when using 'software install' to transfer the Data ONTAP code to the controllers, the 'download' command is also required on both nodes before initiating the takeover process. This has always been necessary for all active/active upgrade operations.
Alternatively, the preferred method is to use the 'software update -r' command, which combines the 'software install' and 'download' commands (the -r switch prevents automatic reboot).
Yes, the first takeover (in your case, as issued from node2) will cleanly shut down and reboot node1. (Halt is not required for the first takeover in a major version NDU.)
As Michael mentioned in step 6 below, when the two nodes are running different major versions of Data ONTAP, active/active functionality is disabled except for the 'cf takeover -n' command issued from node running the higher version of Data ONTAP.
Thanks.
Steve Lawler Technical Marketing Engineer High Availability
NetApp 408.822.1486 steve.lawler@netapp.com www.netapp.com
-----Original Message----- From: Jerry [mailto:juanino@yahoo.com] Sent: Monday, 11 August, 2008 7:23 To: list toasters; Strickland, Michael Subject: RE: 7.0.6 to 7.2.5.1 upgrade
I used the upgrade advisor plans sent by netapp. What is interesting is I never remember in past upgrades doing download on both filers. (is this something new?). I basically did this:
filer1> download (on one filer) filer1> version -b filer2> cf takeover
--- note, filer1 immediately rebooted here, with no option for me to issue a halt ----
as filer1 was coming back up, I noticed filer2 never completed his takeover. once filer1 was back up cf status reported cf disabled because of version mismatch.
So, was filer1 supposed to reboot right away when filer2 did a takeover? I am trying to figure out what I did wrong here, but I think I did all the steps ok. On the update of filer2 I did not have a problem, as I did a cf takeover -n there which allowed me to run halt there.
To answer the other person's post, failover has worked about a week prior so it's a known working cluster.
--- On Sat, 8/9/08, Strickland, Michael Michael.Strickland@netapp.com wrote:
From: Strickland, Michael Michael.Strickland@netapp.com Subject: RE: 7.0.6 to 7.2.5.1 upgrade To: juanino@yahoo.com, "list toasters" toasters@mathworks.com Date: Saturday, August 9, 2008, 2:01 PM That sometimes happens if you are attempting a non-disruptive upgrade and either miss a step, or something (download for example) does not complete as it should.
So, if you decide not to "cf disable", install, download, and reboot both heads individually (typical DISRUPTIVE upgrade), you have to follow steps simliar to the ones below.
- Install software to both nodes
- "download" on both nodes
- Run "version -b" to ensure that both nodes
have the new kernel version on the primary cf partition 4. filer1> cf takeover (filer2 now reboots, and you should notice the new kernel version mentioned. It then waits for giveback) 5. filer1> cf giveback -f (filer2 continues to boot) 6. At this point, filer2 is on the new ontap version, filer1 is downrev (takeover not possible) 7. filer1> halt (then immediately enter the following command on filer2) 8. filer2> cf takeover -n (-n switch used to allow NDU upgrade) 9. allow filer1 to boot and wait for giveback.. Note it should be on the new Ontap kernel at this point. 10. filer2> cf giveback -f 11. Check active version on each filer filer*> version
Note that these are the "basic" steps to perform a non-disruptive upgrade. More of an outline really. You should check the NOW site to ensure that Ontap versions involved are NDU capable, as well as any potential issues if SAN is involved.
Regards,
Michael Strickland Netapp Support
-----Original Message----- From: Jerry [mailto:juanino@yahoo.com] Sent: Friday, August 08, 2008 6:50 PM To: list toasters Subject: 7.0.6 to 7.2.5.1 upgrade
Has anyone experienced an issue going from 7.0.6 to 7.2.5.1. I did an upgrade on my 940 cluster and when I attempted takeover the takeover never succeeded. The first head decided to reboot (without me running halt) but the second head never completed the takeover.
After that it complained about version mismatch.
We're trying to make Snapdrive 6 work on windows 2003 in a vmware guest OS environment with data on FC. We're getting Error 1920 (service won't start). Anyone run into these issues?
Thanks Suresh
Hi,
the same here. We stay with 5.0.
with kind regards ______________________________________creating IT solutions
Dipl.-Chem. Knut Kristan Weber
Senior Solutions Engineer CAx Professional Services
science + computing ag Hagellocher Weg 73 phone +49 7071 94 57 473 72070 Tübingen, Germany fax +49 7071 94 57 411 http://www.science-computing.de
We didn't use software install, we used the NFS install_netapp over NFS to place the files there. If I do this method, I assume running download still required on both heads then. Support has narrowed down my problem to this:
head01/head02: raid.assim.tree.noRootVol:error]: No usable partner root volume found!
They claim this is due to the firmware update of disks still occuring when I did a takeover, but I wonder if it's not because download wasn't run on both systems.??
Thanks for clarification on the takeover/halt/reboot confusion.
--- On Mon, 8/11/08, Lawler, Steve Steve.Lawler@netapp.com wrote:
From: Lawler, Steve Steve.Lawler@netapp.com Subject: RE: 7.0.6 to 7.2.5.1 upgrade To: juanino@yahoo.com, "list toasters" toasters@mathworks.com, "Strickland, Michael" Michael.Strickland@netapp.com Date: Monday, August 11, 2008, 5:37 PM Hi, Jerry...
Yes, when using 'software install' to transfer the Data ONTAP code to the controllers, the 'download' command is also required on both nodes before initiating the takeover process. This has always been necessary for all active/active upgrade operations.
Alternatively, the preferred method is to use the 'software update -r' command, which combines the 'software install' and 'download' commands (the -r switch prevents automatic reboot).
Yes, the first takeover (in your case, as issued from node2) will cleanly shut down and reboot node1. (Halt is not required for the first takeover in a major version NDU.)
As Michael mentioned in step 6 below, when the two nodes are running different major versions of Data ONTAP, active/active functionality is disabled except for the 'cf takeover -n' command issued from node running the higher version of Data ONTAP.
Thanks.
Steve Lawler Technical Marketing Engineer High Availability
NetApp 408.822.1486 steve.lawler@netapp.com www.netapp.com
-----Original Message----- From: Jerry [mailto:juanino@yahoo.com] Sent: Monday, 11 August, 2008 7:23 To: list toasters; Strickland, Michael Subject: RE: 7.0.6 to 7.2.5.1 upgrade
I used the upgrade advisor plans sent by netapp. What is interesting is I never remember in past upgrades doing download on both filers. (is this something new?). I basically did this:
filer1> download (on one filer) filer1> version -b filer2> cf takeover
--- note, filer1 immediately rebooted here, with no option for me to issue a halt ----
as filer1 was coming back up, I noticed filer2 never completed his takeover. once filer1 was back up cf status reported cf disabled because of version mismatch.
So, was filer1 supposed to reboot right away when filer2 did a takeover? I am trying to figure out what I did wrong here, but I think I did all the steps ok. On the update of filer2 I did not have a problem, as I did a cf takeover -n there which allowed me to run halt there.
To answer the other person's post, failover has worked about a week prior so it's a known working cluster.
--- On Sat, 8/9/08, Strickland, Michael Michael.Strickland@netapp.com wrote:
From: Strickland, Michael
Subject: RE: 7.0.6 to 7.2.5.1 upgrade To: juanino@yahoo.com, "list toasters"
Date: Saturday, August 9, 2008, 2:01 PM That sometimes happens if you are attempting a non-disruptive upgrade and either miss a step, or something (download for
example)
does not complete as it should.
So, if you decide not to "cf disable",
install,
download, and reboot both heads individually (typical DISRUPTIVE upgrade),
you
have to follow steps simliar to the ones below.
- Install software to both nodes
- "download" on both nodes
- Run "version -b" to ensure that both
nodes
have the new kernel version on the primary cf partition 4. filer1> cf takeover (filer2 now reboots, and
you
should notice the new kernel version mentioned. It then waits for giveback) 5. filer1> cf giveback -f (filer2 continues to
boot)
- At this point, filer2 is on the new ontap
version,
filer1 is downrev (takeover not possible) 7. filer1> halt (then immediately enter the following command on filer2) 8. filer2> cf takeover -n (-n switch used to
allow
NDU upgrade) 9. allow filer1 to boot and wait for giveback..
Note it
should be on the new Ontap kernel at this point. 10. filer2> cf giveback -f 11. Check active version on each filer filer*>
version
Note that these are the "basic" steps to perform a non-disruptive upgrade. More of an outline really. You should check
the
NOW site to ensure that Ontap versions involved are NDU capable,
as
well as any potential issues if SAN is involved.
Regards,
Michael Strickland Netapp Support
-----Original Message----- From: Jerry [mailto:juanino@yahoo.com] Sent: Friday, August 08, 2008 6:50 PM To: list toasters Subject: 7.0.6 to 7.2.5.1 upgrade
Has anyone experienced an issue going from 7.0.6 to 7.2.5.1. I did an upgrade on my 940 cluster and when I attempted
takeover the
takeover never succeeded. The first head decided to reboot
(without
me running halt) but the second head never completed the
takeover.
After that it complained about version mismatch.
Thanks Steve (and others who responded). My update for my 3050c went great!
The root of my problem was the disk fw update running while performing the takeover. Note to others, do your disk fw before, it takes about 2.5 seconds per disk in my setup.
If you happen to be offlining a disk during the takeover you'll get this: raid.assim.tree.noRootVol:error]: No usable partner root volume found!
--- On Wed, 8/13/08, Jerry juanino@yahoo.com wrote:
From: Jerry juanino@yahoo.com Subject: RE: 7.0.6 to 7.2.5.1 upgrade To: "list toasters" toasters@mathworks.com, "Strickland, Michael" Michael.Strickland@netapp.com, "Lawler, Steve" Steve.Lawler@netapp.com Date: Wednesday, August 13, 2008, 9:50 PM We didn't use software install, we used the NFS install_netapp over NFS to place the files there. If I do this method, I assume running download still required on both heads then. Support has narrowed down my problem to this:
head01/head02: raid.assim.tree.noRootVol:error]: No usable partner root volume found!
They claim this is due to the firmware update of disks still occuring when I did a takeover, but I wonder if it's not because download wasn't run on both systems.??
Thanks for clarification on the takeover/halt/reboot confusion.
--- On Mon, 8/11/08, Lawler, Steve Steve.Lawler@netapp.com wrote:
From: Lawler, Steve Steve.Lawler@netapp.com Subject: RE: 7.0.6 to 7.2.5.1 upgrade To: juanino@yahoo.com, "list toasters"
toasters@mathworks.com, "Strickland, Michael" Michael.Strickland@netapp.com
Date: Monday, August 11, 2008, 5:37 PM Hi, Jerry...
Yes, when using 'software install' to transfer
the
Data ONTAP code to the controllers, the 'download' command is
also
required on both nodes before initiating the takeover process. This has
always
been necessary for all active/active upgrade operations.
Alternatively, the preferred method is to use the 'software update -r' command, which combines the 'software install'
and
'download' commands (the -r switch prevents automatic reboot).
Yes, the first takeover (in your case, as issued from node2) will cleanly shut down and reboot node1. (Halt is not
required
for the first takeover in a major version NDU.)
As Michael mentioned in step 6 below, when the two
nodes
are running different major versions of Data ONTAP, active/active functionality is disabled except for the 'cf takeover -n'
command
issued from node running the higher version of Data ONTAP.
Thanks.
Steve Lawler Technical Marketing Engineer High Availability
NetApp 408.822.1486 steve.lawler@netapp.com www.netapp.com
-----Original Message----- From: Jerry [mailto:juanino@yahoo.com] Sent: Monday, 11 August, 2008 7:23 To: list toasters; Strickland, Michael Subject: RE: 7.0.6 to 7.2.5.1 upgrade
I used the upgrade advisor plans sent by netapp. What
is
interesting is I never remember in past upgrades doing download on
both
filers. (is this something new?). I basically did this:
filer1> download (on one filer) filer1> version -b filer2> cf takeover
--- note, filer1 immediately rebooted here, with no
option
for me to issue a halt ----
as filer1 was coming back up, I noticed filer2 never completed his takeover. once filer1 was back up cf status reported
cf
disabled because of version mismatch.
So, was filer1 supposed to reboot right away when
filer2
did a takeover? I am trying to figure out what I did wrong here, but I think I did all the steps ok. On the update of filer2 I did not have
a
problem, as I did a cf takeover -n there which allowed me to run
halt
there.
To answer the other person's post, failover has
worked
about a week prior so it's a known working cluster.
--- On Sat, 8/9/08, Strickland, Michael Michael.Strickland@netapp.com wrote:
From: Strickland, Michael
Subject: RE: 7.0.6 to 7.2.5.1 upgrade To: juanino@yahoo.com, "list toasters"
Date: Saturday, August 9, 2008, 2:01 PM That sometimes happens if you are attempting a non-disruptive upgrade and either miss a step, or something (download
for
example)
does not complete as it should.
So, if you decide not to "cf disable",
install,
download, and reboot both heads individually (typical DISRUPTIVE
upgrade),
you
have to follow steps simliar to the ones below.
- Install software to both nodes
- "download" on both nodes
- Run "version -b" to ensure that
both
nodes
have the new kernel version on the primary cf partition 4. filer1> cf takeover (filer2 now
reboots, and
you
should notice the new kernel version mentioned. It then waits
for
giveback) 5. filer1> cf giveback -f (filer2 continues
to
boot)
- At this point, filer2 is on the new ontap
version,
filer1 is downrev (takeover not possible) 7. filer1> halt (then immediately enter
the
following command on filer2) 8. filer2> cf takeover -n (-n switch used
to
allow
NDU upgrade) 9. allow filer1 to boot and wait for giveback..
Note it
should be on the new Ontap kernel at this point. 10. filer2> cf giveback -f 11. Check active version on each filer
filer*>
version
Note that these are the "basic" steps
to
perform a non-disruptive upgrade. More of an outline really. You should
check
the
NOW site to ensure that Ontap versions involved are NDU
capable,
as
well as any potential issues if SAN is involved.
Regards,
Michael Strickland Netapp Support
-----Original Message----- From: Jerry [mailto:juanino@yahoo.com] Sent: Friday, August 08, 2008 6:50 PM To: list toasters Subject: 7.0.6 to 7.2.5.1 upgrade
Has anyone experienced an issue going from 7.0.6
to
7.2.5.1. I did an upgrade on my 940 cluster and when I attempted
takeover the
takeover never succeeded. The first head decided to
reboot
(without
me running halt) but the second head never completed the
takeover.
After that it complained about version mismatch.
-
Jerry -
Knut Kristan Weber -
Lawler, Steve -
Raj Patel -
Strickland, Michael -
Suresh Rajagopalan