Ticket #281 (closed task: fixed)

Opened 2 years ago

Last modified 2 years ago

admin page broken on multibox dev server

Reported by: russ Assigned to: ronald
Priority: high Milestone:
Component: topaz-mulgara Version:
Keywords: Cc:
Blocking: Blocked By:

Description

eric is already working on this, but i figured i'd submit a ticket while i'm here.

we need to test rc9 on the multibox development stack before we can push it live.

we have all the servers that used to live at topaz up and running here.

we changed their host names, ip addresses, and config files.

admin/adminTop.action returns a site error and we have no idea why. everything else seems to be working.

you can access this stack at webhead-dev.plos.org plosone-dev topaz-dev fedora-dev mulgara-dev

url is http://webhead-dev.plos.org

when on the PLoS VPN.

Dependency Graph

Change History

02/15/07 14:55:12 changed by ronald

  • owner changed from stevec to ronald.
  • component changed from user-interface to mulgara.

First I assume you mean to test 0.5, as there is no rc9.

Second, the error is due to a bug in Mulgara: part of the database (the string-pool) got corrupted. This is being looked into. Since this is not a bug on our side, and we have a snapshot of the corrupted db files, it might be best to rm the database and start afresh with an empty db so you can proceed with your testing.

I checked production a couple days ago, btw., and there's no sign of this corruption there.

02/16/07 11:48:14 changed by russ

yes. rc-0.5-1, correct.

what's my best path forward here? is there a set of files i can delete or a command i can run to clean out the mulgara database and keep working? or should i blow everything away and install rc8 from scratch and then upgrade?

thanks!

02/16/07 13:46:37 changed by amit

Depends on what you want to test. If a upgrade from RC8 to 0.5, then yes, you will have to blow away both Fedora and Mulgara, run tests with RC8 first to get back to normal and then do an upgrade (I am presuming you are talking about test machines here and not production). If you want to do this the hard way (to get a better understanding of how the system functions) try blowing away the data directories (/var/spool/data or something like that with mulgara and fedora underneath it) or do an install from scratch.

02/16/07 15:45:48 changed by russ

  • status changed from new to closed.
  • resolution set to fixed.

here's what i did:

stop all services

mulgara: rm -fr /var/lib/topaz/data/mulgara

fedora: drop database fedora21 rm -fr /var/lib/topaz/data/fedora reinitialize fedora database

cas: truncate plos_user

start all services.

create admin user and assign admin user.

everything works! sweeeeeeet.