[RAS] Fwd: [ACIS-tech] Migration Help
Ivan Kurmanov
duraley at gmail.com
Tue Aug 30 11:39:10 CDT 2011
---------- Forwarded message ----------
From: Ivan Kurmanov <duraley at gmail.com>
Date: Tue, Aug 30, 2011 at 7:37 PM
Subject: Re: [ACIS-tech] Migration Help
To: ACIS technical list <acis-tech at lists.openlib.org>
On Tue, Aug 30, 2011 at 7:22 PM, Dan Hayes <dhayes501 at gmail.com> wrote:
> It looks like it will take editing /etc/hosts to access the web interface.
> Set authors.repec.org to 65.89.18.160.
>
> Also, the Affiliations section appears to be working fine now.
I witness.
Which means the Storable strings that were written (with nfreeze) on
the old nebka are readable on the new one, and we do not need any
other serialization changes at the moment.
>>> Dan, did you measure or record the time it took to restore the database?
>> It took a solid 8 hours. I didn't time it exactly though. I watched the
>> size of /var/lib/mysql and it grew to over 35G, acis is 32G and rdb is 4.2G
Then it looks like we should plan for at least 10-12 hours of downtime
of the service for the final migration.
>From the point of view of the least user experience damage (and brand
damage), it would be great to do the migration over the weekend. But
the nearest weekend (3-4 Sep) i'm occupied except at night, when I'd
like to have some sleep. The weekend after (10-11 Sep) i'm available.
Or we can ignore the weekends and plan the transition on a business
day. I guess Dan will do the large part of the work again, so his
availability is key here. I'm pretty flexible at the moment.
And of course, there's more tests and checks to be done.
-i
>
> On Tue, Aug 30, 2011 at 10:19 AM, Dan Hayes <dhayes501 at gmail.com> wrote:
>>
>> It took a solid 8 hours. I didn't time it exactly though. I watched the
>> size of /var/lib/mysql and it grew to over 35G, acis is 32G and rdb is 4.2G
>>
>> On Tue, Aug 30, 2011 at 10:05 AM, Ivan Kurmanov <duraley at gmail.com> wrote:
>>>
>>> Dan, did you measure or record the time it took to restore the database?
>>>
>>> -ivan
>>>
>>> On Tue, Aug 30, 2011 at 5:41 PM, Dan Hayes <dhayes501 at gmail.com> wrote:
>>> > The SQL database has been loaded, and it appears to be running
>>> > properly.
>>> > You can test it at 65.89.18.160 or at http://nebka.stlouisfed.org.
>>> >
>>> > Did I not need to run the updareq commands again? It appears to be
>>> > working
>>> > without me needing to do anything else.
>>> >
>>> > On Mon, Aug 29, 2011 at 10:16 AM, Dan Hayes <dhayes501 at gmail.com>
>>> > wrote:
>>> >>
>>> >> FYI
>>> >>
>>> >> I rsynced the files Saturday afternoon, and dumped the sql database
>>> >> then
>>> >> as well. I transferred the db backup on Sunday, and I'm now restoring
>>> >> the
>>> >> sql databases on the STL server. It looks like its going to take a
>>> >> while.
>>> >>
>>> >> On Sun, Aug 28, 2011 at 9:22 AM, Thomas Krichel <krichel at openlib.org>
>>> >> wrote:
>>> >>>
>>> >>> Dan Hayes writes
>>> >>>
>>> >>> > So the reason is that its faster than MySQL for what its being used
>>> >>> > for?
>>> >>>
>>> >>> Yes. At least we think so.
>>> >>>
>>> >>> Kindly send me an ssh public key to give you access to the test
>>> >>> server.
>>> >>>
>>> >>>
>>> >>> Cheers,
>>> >>>
>>> >>> Thomas Krichel http://openlib.org/home/krichel
>>> >>> http://authorprofile.org/pkr1
>>> >>> skype: thomaskrichel
>>> >>>
>>> >>> _______________________________________________
>>> >>> ACIS-tech mailing list
>>> >>> ACIS-tech at lists.openlib.org
>>> >>> http://lists.openlib.org/mailman/listinfo/acis-tech
>>> >>
>>> >
>>> >
>>> > _______________________________________________
>>> > ACIS-tech mailing list
>>> > ACIS-tech at lists.openlib.org
>>> > http://lists.openlib.org/mailman/listinfo/acis-tech
>>> >
>>>
>>> _______________________________________________
>>> ACIS-tech mailing list
>>> ACIS-tech at lists.openlib.org
>>> http://lists.openlib.org/mailman/listinfo/acis-tech
>
>
> _______________________________________________
> ACIS-tech mailing list
> ACIS-tech at lists.openlib.org
> http://lists.openlib.org/mailman/listinfo/acis-tech
>
More information about the RAS-run
mailing list