Hello all,
We are going to be performing a somewhat emergency upgrade on our
Magnolia 3.0.2 instance next week because we are experiencing a couple
of database issues that are no longer acceptable to the client:
1. As I have mailed about several times, but not really received any
suggestions as to how to remedy, our database has grown out of
control. Our 50-60 page site has an authoring database over
1.2 GB at this time, and the public one is over 300 MB. The sheer
size of these databases is being caused almost entirely by the
version tables. The site has only been in operation for 8 months
and sees fairly low editing activity most weeks, yet grows several
hundred MBs a month. The nightly backups for these are chewing up
disk space like candy.
2. The client is experiencing extremely long lag times from activating
a page to seeing it in the Inbox, and more often than we would like
experiences corruption in the workflow process
(javax.jcr.RepositoryException: failed to retrieve item state of
item...) that requires us to drop the Expressions and Store tables
to clear it up before they are able to perform any page activations
again.
We have tried to rid ourselves unsuccessfully of these enormous version
tables by following the instructions on the Magnolia documentation
website for disabling versioning, thinking that once the versioning was
turned off we could drop the versioning tables. This was unfortunately
a bad idea and put our database in a bad state where nothing could be
activated because there were missing nodes (presumably from the version
tables we dropped... whoops), and we had to restore from a backup.
We have decided to export our existing repositories to XML and hopefully
rid ourselves of the workflow and versioning entirely since these are
apparently disabled by default in 3.5.3. Our client has requested that
we remove the workflow/versioning because they are causing more trouble
than they are worth to the client. However, now we are faced with
another problem that I've unsuccessfully in the past asked for advice
about... the export process.
I have never been able to successfully obtain an XML export from our
site from the admin Tools -> Export page. We are able to export
pages/page trees from the Website view fine, but whenever we use the
Export page:
1. It takes hours to produce an export file that is only about 2-3
MBs in size. I tried to generate one on Friday on my local test
server and had to kill the process after 3 hours without a file
being produced. I am assuming, but have no real idea, that this
is related to our 1.2 GB database size, even though I have not
requested any version information be kept. The last time I tried
to get an export, it took about 1.5 hours but at the time the
database was about half the size.
2. Once we do obtain the XML export, all the line breaks are
converted to <br> which produces completely messed up pages on
import. In the past some users suggested we make sure that we
do not have formatting selected as an export option, but we do
always leave this blank and the XML is still formatted. I have
seen several mentions in JIRA that something like this was
supposedly fixed in 3.0 Final, but this is 3.0.2 we see this on.
Right now my only option is really to go through the Website view and
export each page tree individually, but I'd really like to do it the
"right" way and get the whole repository at once. Does anyone have any
suggestions on how we could actually export the whole website repository
the correct way? Anything I can do to speed it up or make it be
formatted correctly? If we upgrade to 3.0.5 first, will that help the
export speed and/or the formatting problems? I am afraid to run the
export process on the production site right now and possibly cause
performance and/or memory issues.
Thank you in advance for any advice on our exports!
-- Miranda
--
Miranda Jones
Objective Consulting, Inc.
http://www.spiders.com
----------------------------------------------------------------
for list details see
http://documentation.magnolia.info/docs/en/editor/stayupdated.html
----------------------------------------------------------------
