[Gluster-users] Errors in documentation for 3.2 to 3.3 upgrade path

Toby Corkindale toby.corkindale at strategicdata.com.au
Thu Jan 24 01:29:38 UTC 2013


On 24/01/13 10:57, Joe Julian wrote:
> On 01/23/2013 03:43 PM, Toby Corkindale wrote:
>> Hi,
>> Last night I attempted to upgrade some GlusterFS servers from 3.2.x to
>> 3.3.1.
>>
>> The upgrade did NOT go smoothly, and I'm quite disappointed in the
>> documentation for the upgrade as it was quite erroneous.
>>
>> I followed this guide:
>> http://www.gluster.org/2012/05/upgrading-to-glusterfs-3-3-0/
>>
>> The part I take issue with is step 5, which mentions:
>>  glusterd –xlator-option *.upgrade=on -N
>>
>> That command very definitely does not run successfully.
>> Reading the docs and the comments made it clear that a hypen was
>> missing, however even the following command only returns an error:
>>
>>  glusterd --xlator-option *.upgrade=on -N
>>
>> I then took a guess based on the docs, and tried this instead:
>>
>>  glusterd --xlator-option*.upgrade=on -N
>>
>> That command exited without an error message, so initially I thought
>> it had succeeded; however when I checked the contents of the vol
>> files, I found that they had not been changed or upgraded at all.
>> In /var/log/gluster/ there was a file called
>> "--xlator-option*.upgrade=on.log"
>> and it just contained errors about invalid mount points.
>>
>>
>> Sooooooooo my questions are:
>> 1) what is the correct way to upgrade a 3.2 cluster to 3.3?
>> 2) What are the consequences of having effectively skipped whatever
>> that command was meant to do?
>> 3) how can an eight month old wiki page on the upgrade be so wrong and
>> yet no-one has noticed??
>>
> 1) From the RPM scripts, I copied:
>
> glusterd --xlator-option *.upgrade=on -N
>
> You didn't specify what error you noticed, so I can't go any further
> with that question.

I had a look at my bash history, and it appears that the command I was 
entering was using two hypens, but the second one was the unicode hypen 
and the first was the keyboard hyphen. (ie. I had copied and pasted, 
then added an extra hyphen). That'll explain the errors I was seeing.

Re-running the command, as copied from your email above, does seem to 
work fine, and does modify the .vol files in /var/lib/glusterd

> 2) The consequences I pulled from that same "wiki" page (actually it's a
> wordpress article), "This will re-generate volume files with the new
> ‘index’ translator which is needed for features like pro-active self
> heal in 3.3.0."

Ah OK, I was wondering what happens if you don't have that stuff.. It 
sounds like the self-healing will be broken, but not other stuff?



> 3) I suspect the version migration from the former wordpress to the
> current did that stupid hyphen replacement thing that plagues many tech
> articles all over the internet. For some reason that software likes to
> convert "--" to "&hyph;". Another reason that nobody's complained may be
> that the RPM packaging does that for you so it's superfluous to those users.
>
> That should get fixed though. Thanks for pointing it out.

Thanks for dealing with the bug report quickly and helpfully.

-Toby




More information about the Gluster-users mailing list