DokuWiki

Save the file attached to the ticket 6718

save it as ghosttest.c in the target machine

gcc ghosttest.c -o ghosttest ./ghosttest

output should be “not vulnerable” If it says vulnable, then we have to

yum update glibc* and then again validate with ./ghosttest

If you run yum update glibc - we need to reboot the servers.

Apply the patches above and reboot should be done in this sequence - please change and validate order from Uday

(1) prdapp02 - (No down time is needed).

  before restart
   -> use internal IP, validate site.
   -> check running services and verify that everything is normal
   -> verify whether bash update was done. if so verify whether it will impact restart -Opsource ticket 42400
   -> update network functions if the above scenario is valid
   -> restart prdapp02 
  after restart
   -> Verify that the services are running - (apache/passenger, god & gluster drive is mounted. check if rails console opens up normally and no side effects.)
   -> check the site with internal IP
   -> check the site with external IP
   -> confirm normal functioning of site

(2) mysql slave - (Down time is needed (because of content files)).

  before restart
   -> check the status of slave with log and position replicated . check whether Yes is there for mysql slave replication process and thread 
   -> stop slave mysql
   -> verify whether bash update was done. if so verify whether it will impact restart -Opsource ticket 42400
   -> update network functions if the above scenario is valid
   -> restart the box
  after restart
    -> check mysql services, solr and gluster server. Verify that gluster mounts are fine on prodapp01 and prodapp02. 
    -> start slave mysql
    -> verify that the replication has started without errors- show slave status\G 

(3) mysql master - (Down time is needed).

  before restart
   -> check running services
   -> verify whether bash update was done. if so verify whether it will impact restart -Opsource ticket 42400
   -> update network functions if the above scenario is valid
   -> restart master
  after restart
   -> check mysql services
   -> show master status

(4) prdapp01 - (Down time is needed).

  before restart
   -> use internal IP, validate site.
   -> check running services and verify that everything is normal
   -> verify whether bash update was done. if so verify whether it will impact restart -Opsource ticket 42400
   -> update network functions if the above scenario is valid
   -> restart prdapp01
  after restart
   -> Verify that the services are running - (apache/passenger, god & gluster drive is mounted. check if rails console opens up normally and no side effects.)
   -> check the site with internal IP
   -> check the site with external IP
   -> confirm normal functioning of site

(5) prodcollab01 - (Down time is needed because of video recording feature.)

  before restart
  -> check gluster services running
  -> check file / folder ownerships and document them for the gluster service 
  -> raise a ticket for Opsource that you are going to restart the box
  -> be on call for them connected to the terminal of the box
  -> restart and then ask them to skip disk errors as it will take hours for the disk errors to be checked
  after restart
  -> check bbb / gluster services
  -> check gluster mounts from prdapp01 / prdapp02

(6) front end - (Down time is needed)

  before restart
  after restart
  -> Verify Haproxy is up. Chat service is fine (both redis and node).

After the complete reboot -

Issue 1 - prodcollab01 - call Opsource and request them to skip disk check as we reboot Issue 2 - Network restart - https://bugzilla.redhat.com/show_bug.cgi?id=482826

Follow the solution in the bug above before doing the restart on each and every system.