[ubuntu-cloud] UEC upgrade 1.6.2-2.0.1, problems with HSQLDB
Alexander Novikov
novikov at wdcb.ru
Mon Dec 13 14:42:54 GMT 2010
Hello everyone!
I`ve got problem during automated upgrade process from Eucalyptus 1.6.2 to
2.0.1 on UEC x64 10.0.4LTS to Ubuntu 10.10 server.
In two words - apt-get and other tools(update-manager) find&download and
replace existing packages with new ones. With some packages there are
configuration changes - so process ask user what config will be - old or new.
So, the problem.
I`ve replace /etc/eucaliptus/eucalyptus.conf ,cause some new values added, and
remove JVM_MEM option "512m";. But when started some process that reconfig
HSQLDB eucaliptus tables, it faced with java heap error, try to reconnect &
reconfig, but finally stuck. In some time, after manually
running " --configure eucalyptus-java-common" and removing and then
installing it again, DB reconfig process was also not successed, but it ends
his work (not with Ctrl+C key) and update somehow finished.
As it obvious in such condition there is now working cloud HSQLDB, but I also
has some previous backups.
My question is - can I somehow use these backups to restore previous 1.6 Cloud
configuration (Users,Images, etc) and will I have to reconfig it again? If
so - what commands can you give in advice?
Further ---------
Has anyone some ideas where I`m mistaken?
On Ubuntu 10.10 server I temporary change apt repos to LTS and install
eucalyptus v.1.6.2. Than recreate or check /var/lib/eucalyptus
(especially ./keys ./.ssh and ./db directories ), than change back repos and
do upgrade - only eucalyptus packages upgraded. Again error with java heap
space - possible walrus db is too large:
root at cloud:~# ls -la /var/lib/eucalyptus/db/ total 40344 -rw-r--r-- 1
eucalyptus eucalyptus 41093528 2010-12-10 21:00 eucalyptus_walrus.script
Now, there are such errors as - downgraded eucalyptus 1.6.2 doesnt work
correctly and even allow to configure itself (web interface doesnt work,
euca_conf --register-cluster .. also). Upgraded finally lost all db values
and go with a new one made by own.
Another addition,
finally I`ve managed to reduce DB size for storage, so script somehow
successeded, but again, when terminating, - java_heap_space:
[Server at 54135640]: Initiating shutdown sequence...
Exception in thread "HSQLDB Connection @3126cb1a"
java.lang.NullPointerException
at org.hsqldb.ServerConnection.close(Unknown Source)
at org.hsqldb.ServerConnection.run(Unknown Source)
at java.lang.Thread.run(Thread.java:619)
Exception in thread "HSQLDB Connection @75cfa3d2"
java.lang.NullPointerException
at org.hsqldb.ServerConnection.close(Unknown Source)
at org.hsqldb.ServerConnection.run(Unknown Source)
at java.lang.Thread.run(Thread.java:619)
Exception in thread "HSQLDB Connection @1f2a45c7"
java.lang.NullPointerException
at org.hsqldb.ServerConnection.close(Unknown Source)
at org.hsqldb.ServerConnection.run(Unknown Source)
at java.lang.Thread.run(Thread.java:619)
Exception in thread "HSQLDB Connection @5ba28182"
java.lang.NullPointerException
at org.hsqldb.ServerConnection.close(Unknown Source)
at org.hsqldb.ServerConnection.run(Unknown Source)
at java.lang.Thread.run(Thread.java:619)
Exception in thread "HSQLDB Connection @ef0cda1"
java.lang.NullPointerException
at org.hsqldb.ServerConnection.close(Unknown Source)
at org.hsqldb.ServerConnection.run(Unknown Source)
at java.lang.Thread.run(Thread.java:619)
17:06:37 ERROR 98 StandalonePersistence | java.lang.OutOfMemoryError: Java
heap space
Exception in thread "HSQLDB Connection @48da7565"
java.lang.NullPointerException
All important bukkits alive, but now there are errors in cloud-controller due
to its not first-time_configuration but already existing storage pool...
Please help, what steps will recover this cloud?
What should I do to use data from old HSLDB?
Where should I put options for JVM to not overheaped? (in eucalyptus.conf
looks like it doesnt matter)
or another comments, please
Thank you for your attention!
More information about the Ubuntu-cloud
mailing list