eon upgrade question

[vajonam]

dre2kse,

I have had eon running for sometime, and really haven't touched it, i am now upgrading a case and adding some hot swap bays etc, and thought it be a good time to upgrade the base image to the latest. 

SunOS SEVERNAME 5.11 oi_151a i86pc i386 i86p

is what I am running. 

1. exec transporter.sh -i /path/to.iso -b /path/to_backup/ -d /mnt/eon0
1. as I understand this will copy the contents of the iso including the new image as well as the .backup and .exec files.
2. then i need to re-edit the .backup and .exec files to represent any changes i have made in my system, can i use the old .backup and .exec files instead, really transporter shouldn't be overwriting my configuration should it? 
2. run updimg.sh /mnt/eon0, to make sure my files including mpt_sas driver etc are copied over to the new image, making sure the exclude the setup, install.sh and updimg.sh from the .backup file and re-adding it later? 
3. cross fingers and re-boot.
   

Can you confirm that I have it right?

thanks

 

[Andre Lue]

Yes, that sounds right. Good to see you still around Vajonam. I would skip/comment certain files in .backup (for example don't want to copy older /bin/rail over the new one, /etc/.bashrc etc). Or use the newer versions where possible. Everything else seems correct.

Give the new pkgsrc binkit a shot also(15000+ pkgs). Move your old binkit to a backup location so you can roll backup if there are any showstoppers.

regards

a

--
You received this message because you are subscribed to the Google Groups "EON ZFS Storage" group.
To unsubscribe from this group and stop receiving emails from it, send an email to eonstorage+unsubscribe@googlegroups.com.
To post to this group, send email to eonst...@googlegroups.com.
Visit this group at https://groups.google.com/group/eonstorage.
For more options, visit https://groups.google.com/d/optout.

[vajonam]

Cool. Thanks.

Re: new binkit not sure where to get it, do I just use the new rail install binkit on my first reboot with the new image? looks like things have changed and I haven't kept up..

On Wednesday, May 24, 2017 at 6:08:44 PM UTC-4, dre2kse wrote:

Yes, that sounds right. Good to see you still around Vajonam. I would skip/comment certain files in .backup (for example don't want to copy older /bin/rail over the new one, /etc/.bashrc etc). Or use the newer versions where possible. Everything else seems correct.

Give the new pkgsrc binkit a shot also(15000+ pkgs). Move your old binkit to a backup location so you can roll backup if there are any showstoppers.

regards

a

On Wed, May 24, 2017 at 4:44 PM, vajonam via EON ZFS Storage <eonstorage+APn2wQeaGveeVtAFGo6StkzOM-yyat2Qk_yK0TpsH4zA9NaHLqHYt@googlegroups.com> wrote:

dre2kse,

I have had eon running for sometime, and really haven't touched it, i am now upgrading a case and adding some hot swap bays etc, and thought it be a good time to upgrade the base image to the latest. 

SunOS SEVERNAME 5.11 oi_151a i86pc i386 i86p

is what I am running. 

1. exec transporter.sh -i /path/to.iso -b /path/to_backup/ -d /mnt/eon0
1. as I understand this will copy the contents of the iso including the new image as well as the .backup and .exec files.
2. then i need to re-edit the .backup and .exec files to represent any changes i have made in my system, can i use the old .backup and .exec files instead, really transporter shouldn't be overwriting my configuration should it? 
2. run updimg.sh /mnt/eon0, to make sure my files including mpt_sas driver etc are copied over to the new image, making sure the exclude the setup, install.sh and updimg.sh from the .backup file and re-adding it later? 
3. cross fingers and re-boot.
   

Can you confirm that I have it right?

thanks

 

--
You received this message because you are subscribed to the Google Groups "EON ZFS Storage" group.

To unsubscribe from this group and stop receiving emails from it, send an email to eonstorage+...@googlegroups.com.

[Andre Lue]

See the link for howto http://eonstorage.blogspot.com/2016/11/eon-binary-kit-upgrade-to-pkgsrc_11.html?m=1

The new rail will fetch the pkgsrc bootstrap.

Let me know if you run into issues

To unsubscribe from this group and stop receiving emails from it, send an email to eonstorage+unsubscribe@googlegroups.com.

[vajonam]

Disaster!

I rebuilt my case moving all my HD's and stuff over to a new Norco RPC-2224 case. And now nothing really works. One pool wont import it says invalid vdev configuration, not sure what changed. 

I did move all the HD's over from SATA ports to SAS controller. Before I has some one SATA, some on 3 controllers. 

Other pool says one more more devices are missing.. not sure why it can't see the device! format -e or hdadm display all seems to show me all the devices!  I am f-ed! 

On Wednesday, May 24, 2017 at 10:09:12 PM UTC-4, dre2kse wrote:

See the link for howto http://eonstorage.blogspot.com/2016/11/eon-binary-kit-upgrade-to-pkgsrc_11.html?m=1

The new rail will fetch the pkgsrc bootstrap.

Let me know if you run into issues

[Andre Lue]

Did the pool have any forced/custom/weird zpool setup or mismatched drive size mix to begin with?

Did you boot new hardware with new build, that produced the errors? If yes I would upgrade key n make sure it boot n imports on old hw first. Do you have the old key to boot on the old hw?

What does zdb -l /dev/dsk/cxtxdxsx for each drive show?

What does zpool status show?

To unsubscribe from this group and stop receiving emails from it, send an email to eonstorage+unsubscribe@googlegroups.com.

[vajonam]

I haven't upgraded the software, same hardware just moved some drives around which I thought was going to be okay. but I guess ZFS isn't that resilient to hardware changes after all.. am I wrong. 

Going back to older cables and stuff I have gotten 5 of my 6 pools back. The last one is being stubron and wont import, with a cannot import 'mojave': one or more devices is currently unavailable

Here is the pastebin for one of the drives that's part of the pool called mojave, https://pastebin.com/L6tmB6hX

I am now going to try and move those drives back to the controller in which they were attached to. 

On Thursday, May 25, 2017 at 10:05:32 PM UTC-4, dre2kse wrote:

Did the pool have any forced/custom/weird zpool setup or mismatched drive size mix to begin with?

Did you boot new hardware with new build, that produced the errors? If yes I would upgrade key n make sure it boot n imports on old hw first. Do you have the old key to boot on the old hw?

What does zdb -l /dev/dsk/cxtxdxsx for each drive show?

What does zpool status show?

[vajonam]

Okay that worked. 

I put each pool back on the controller it was on to begin with seems to work. My issues as my system grew, i first started with SATA, then one 8port raid card and then one more. So the newer pools were on the nice fast card while the older ones was on the slower cards. I wanted to move as many pools to the newer hardware as possible. 

So the question is should I be able to move disks around various controllers on the same system? if so I do an export before I import it back? 

Thanks for your pointers! and just confidence in the system. 

[Andre Lue]

ZFS should be controller independent of hw n controller, position once you have all the disk members. Not sure, what controller is label 2, 3 on from your pastebin , they could not be read which prob caused your issue?

On May 25, 2017 10:51 PM, "vajonam via EON ZFS Storage" <eonstorage+APn2wQeaGveeVtAFGo6St...@googlegroups.com> wrote:
>
> Okay that worked. 
>
> I put each pool back on the controller it was on to begin with seems to work. My issues as my system grew, i first started with SATA, then one 8port raid card and then one more. So the newer pools were on the nice fast card while the older ones was on the slower cards. I wanted to move as many pools to the newer hardware as possible. 
>
> So the question is should I be able to move disks around various controllers on the same system? if so I do an export before I import it back? 

>
> Thanks for your pointers! and just confidence in the system. 
>
>
> On Thursday, May 25, 2017 at 10:38:51 PM UTC-4, vajonam wrote:
>>
>> I haven't upgraded the software, same hardware just moved some drives around which I thought was going to be okay. but I guess ZFS isn't that resilient to hardware changes after all.. am I wrong. 
>>
>> Going back to older cables and stuff I have gotten 5 of my 6 pools back. The last one is being stubron and wont import, with a cannot import 'mojave': one or more devices is currently unavailable
>>
>> Here is the pastebin for one of the drives that's part of the pool called mojave, https://pastebin.com/L6tmB6hX
>>
>> I am now going to try and move those drives back to the controller in which they were attached to. 
>>
>>
>>
>>
>>
>>
>> On Thursday, May 25, 2017 at 10:05:32 PM UTC-4, dre2kse wrote:
>>>
>>> Did the pool have any forced/custom/weird zpool setup or mismatched drive size mix to begin with?
>>>
>>> Did you boot new hardware with new build, that produced the errors? If yes I would upgrade key n make sure it boot n imports on old hw first. Do you have the old key to boot on the old hw?
>>>
>>> What does zdb -l /dev/dsk/cxtxdxsx for each drive show?
>>>
>>> What does zpool status show?
>>>

[vajonam]

Here are all my disks https://pastebin.com/Hf3Z9YWg

Here are my controllers https://pastebin.com/fhPpWfCp

Possibly the mpt_sas2 drivers naming is messing with things. as it goes crom cXtXdX to cXtXXXXXXXXXXXXXXXXXXdX  type name. 

On Thursday, May 25, 2017 at 10:59:48 PM UTC-4, dre2kse wrote:

ZFS should be controller independent of hw n controller, position once you have all the disk members. Not sure, what controller is label 2, 3 on from your pastebin , they could not be read which prob caused your issue?

> To unsubscribe from this group and stop receiving emails from it, send an email to eonstorage...@googlegroups.com.
> To post to this group, send email to eons...@googlegroups.com.

[vajonam]

Did the upgrade, it came back but didn't do any of the backups that I needed, also the backup I location is specified isn't mountable because the mpt_sas driver is not loaded... where I can I get that from now? none of the links have it? I need the x64 version of them mpt_sas diver please if you can have a link that would be awsome!

[vajonam]

Also something happened to root password, think I have to roll back to a backup but stilll need the mpt_sas driver.. trying to move the drives to mpt controller instead....aargh!

[vajonam]

when i move the two disks from the mpt_sas to the mpt controller and do an import i get a kernel oops and reboots, i cant capture the output.. now i am in some serious trouble i think.. 

[vajonam]

I copied the mpt_sas from the Hipster release..

This is what I am seeing on bootup, mpt_sas is no longer loading anything. 

kobj_load_module: mpt_sas read header failed

[Manojav Sridhar]

May 29 19:31:11 eon genunix: [ID 538349 kern.notice] kobj_read_file: z_uncompress failed, error code : 0xfffffffb

May 29 19:31:11 eon genunix: [ID 689466 kern.notice] kobj_load_module: mpt_sas read header failed

--
You received this message because you are subscribed to a topic in the Google Groups "EON ZFS Storage" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/eonstorage/NgnIoX2LgKU/unsubscribe.
To unsubscribe from this group and all its topics, send an email to eonstorage+unsubscribe@googlegroups.com.

[vajonam]

I was able to import the pool using the onboard sata, which is good allowed me roll back with transporter.sh! so I am back in biz with the old stick. 

i was able to do a clean move, this time editing the .backup after transporter.sh and saving the mpt_sas drivers. but I noticed a number of issues.

• all the pools imported, but only 2 pools maintained the sharenfs=on, which was lost again on reboot. not sure what is going on there.
  
• rail is broken and cant download bin kit or get my mysql running.

please let know how fix these? 

On Monday, May 29, 2017 at 3:32:43 PM UTC-4, vajonam wrote:

May 29 19:31:11 eon genunix: [ID 538349 kern.notice] kobj_read_file: z_uncompress failed, error code : 0xfffffffb

May 29 19:31:11 eon genunix: [ID 689466 kern.notice] kobj_load_module: mpt_sas read header failed

[Andre Lue]

To avoid driver incompatibility issues, it's best if you re-use/copy the driver from your older version which is also OI bloodline. You will encounter errors as above if its linked against something different.

To unsubscribe from this group and stop receiving emails from it, send an email to eonstorage+unsubscribe@googlegroups.com.

[vajonam]

Any thoughts on the missing zfs sharenfs=on mounts on certain pools? 

On Monday, May 29, 2017 at 4:54:14 PM UTC-4, dre2kse wrote:

To avoid driver incompatibility issues, it's best if you re-use/copy the driver from your older version which is also OI bloodline. You will encounter errors as above if its linked against something different.

To unsubscribe from this group and all its topics, send an email to eonstorage+...@googlegroups.com.

To post to this group, send email to eonst...@googlegroups.com.
Visit this group at https://groups.google.com/group/eonstorage.
For more options, visit https://groups.google.com/d/optout.

[Andre Lue]

No. That info should be kept on disk members in the metadata as far as I know. Since you mention its shows sometimes, I suspect something else may be going on.

Did you upgrade pool and zfs versions with each upgrade of EON or are you running the initial zpool and zfs version?

To unsubscribe from this group and stop receiving emails from it, send an email to eonstorage+unsubscribe@googlegroups.com.

[Manojav Sridhar]

I haven't upgraded the pool since oi_151a. I just did an import to see that everything works, wasn't ready to do the upgrade.. have too much to loose at the moment. 

On Mon, May 29, 2017 at 5:20 PM, EON ZFS Storage on behalf of dre2kse <eonst...@googlegroups.com> wrote:

No. That info should be kept on disk members in the metadata as far as I know. Since you mention its shows sometimes, I suspect something else may be going on.

Did you upgrade pool and zfs versions with each upgrade of EON or are you running the initial zpool and zfs version?

On Mon, May 29, 2017 at 5:13 PM, vajonam via EON ZFS Storage <eonstorage+APn2wQeaGveeVtAFGo6StkzOM-yyat2Qk_yK0TpsH4zA9NaHLqHYt@googlegroups.com> wrote:

Any thoughts on the missing zfs sharenfs=on mounts on certain pools? 

On Monday, May 29, 2017 at 4:54:14 PM UTC-4, dre2kse wrote:

To avoid driver incompatibility issues, it's best if you re-use/copy the driver from your older version which is also OI bloodline. You will encounter errors as above if its linked against something different.

To unsubscribe from this group and all its topics, send an email to eonstorage+unsubscribe@googlegroups.com.

[Andre Lue]

That's what I suspected. It's possibly related to something there. I would re-issue the sharenfs commands and wait until you have stabilized everything before doing any zpool or zfs upgrades.

[vajonam]

Yup okay, I will do that. I am still having some issues with the binkit install That is my zdb -C, https://pastebin.com/DJYNVCcQ

So your suggest I just do the sharenfs on reboot for now, and once I have all the other bits try the zpool upgrade?  

[Andre Lue]

Don't do the upgrade until you have stabilized eon on the new hardware and running stable for a while.

Paste the output or errors back to the rail thread

To unsubscribe from this group and stop receiving emails from it, send an email to eonstorage+unsubscribe@googlegroups.com.