D&C GLug - Home Page

[ Date Index ] [ Thread Index ] [ <= Previous by date / thread ] [ Next by date / thread => ]

Re: [LUG] Fwd: Debian 11 "bullseye" released

 

On Sunday, 15 August 2021 06:36:41 BST Paul Sutton via list wrote:
>
> Debian 11 "bullseye" released                           press@xxxxxxxxxx

Just upgraded my desktop box. 

No major issues (so far!).

It stalled and needed manual rerun of the "dist-upgrade" because "/boot" is 
too small if I don't clear out unneeded old kernels before installing new 
ones. Not sure why I made "/boot" so small, I should probably grow it. But 
cleaning out old stuff before upgrades makes sense (but reboot before upgrade 
if you do).

KDE ran the Akonadi Migration Agent for 5 hours before I tried "akonadictl 
stop" ;) Ran an fsck using akonadictl which found various dross, then a 
vacuum. I was vaguely aware the Akonadi database wasn't 100%, but it wasn't 
causing any obvious issues in use.

Python upgrades, so all the virtualenvs are toast. I just moved the old 
virtualenvs out of the way and recreated from requirements.txt cleanly. The 
internet is awash with tools and scripts to avoid recreating from 
requirements.txt but that seems an over elaboration to me. I expect it is the 
mega dependencies that drag in lots of C code that hurt. Also nearly 
everything seemed still to be in various caches so it was pretty swift.

Managed to overwrite a database name change in a config file by accepting the 
installer version. So wondered where a load of test data had gone till I diffed 
the old and new config files again and gave it more attention than during the 
upgrade and went "doh". 

The default behaviour of "neo4j" is not to upgrade the database when a new 
version of the database is running against the database files. Which is 
probably a wise default for enterprise databases (I've seen automatic upgrades 
like this break very different production databases), but for me just meant I 
had to change the "allow upgrades" from False to a True in neo4j.conf and 
restart the service (after reading the helpful message in the log file, telling 
me why it didn't upgrade my database , +10 for good error messages neo4j).

That said all this container stuff has me thinking "upgrading" is probably the 
wrong approach, too much pet and not enough cow in my desktop. But there is a 
fair bit of churn happening elsewhere in my IT, so for now I'm happy it all 
just works.

 Simon



-- 
The Mailing List for the Devon & Cornwall LUG
https://mailman.dcglug.org.uk/listinfo/list
FAQ: http://www.dcglug.org.uk/listfaq