Server outage.

9 views
Skip to first unread message

Ross Scanlon

unread,
Aug 17, 2022, 10:46:03 PMAug 17
to osm-...@googlegroups.com
Hi All,

It's happened again.  NBN NTD is destroyed by lightning storm.

Earliest I can get a NBN technician to come and replace it is Tuesday
23rd AEDT (GMT +10).

Nothing I can do as they will not allow us to keep a replacement on hand.

Cheers

Ross


Ross Scanlon

unread,
Aug 23, 2022, 6:19:16 PMAug 23
to osm-...@googlegroups.com
Hi All,

Update on the server.

NBN NTD was replaced.

Server still not available as the router, while working perfectly for
wireless connections, will not connect to wired connections.

So I have to get a new router.   At this stage 7-10 days.

Cheers

Ross

Tim Sheerman-Chase

unread,
Aug 26, 2022, 4:53:59 PMAug 26
to osm-...@googlegroups.com
Hi all,

I've created o5m and OsmAnd extracts for UK, Ireland and Australia.
https://archive.org/details/fosm-extracts-in-osmand-obf-format-aug-2022

The planet dump has also been archived here
https://archive.org/details/fosm-planet_20220801013701.o5m

Tim

Ross Scanlon

unread,
Sep 6, 2022, 8:11:13 PMSep 6
to osm-...@googlegroups.com
Hi All,

Latest with the server.

I've had the router replaced but the server is still not connecting via
wired network to it.

Actually everything that was connected via wired network when the
lightning storm went through will not connect to the router.

Router works fine with anything that was not connected at the time of
the storm and via wireless.

So looks like I'm up for a new motherboard for the server as it's built
in network connection or add a network card into it or a wireless
connection.

Unfortunately I'm not close by at the moment and I don't have anyone
with the level of experience to change or add a network card that can do
it close by.

I thought I had found someone but as soon as I said it was a server
running linux and needed new hardware installed they totally lost it. 
"Only know Windoze and not prepared to learn anything else."

So end of a long story looks like it will be out of action until I'm
back there beginning of November.

Not much else I can do at the moment.  Always seems to be that these
things happen when I'm away and it runs perfectly when I'm there.


Cheers

Ross

Tim Sheerman-Chase

unread,
Oct 3, 2022, 12:11:55 PMOct 3
to osm-...@googlegroups.com
I misread this and thought you said October. That's so long! :) My edit
backlog is getting rather big.

I think we need to revisit how the project is hosted but not in the
sense of removing it to another hosting. We need some solution that
distributes and replicates between postgres instances. Perhaps we can
use postgres built in replication or some sort of middleware between the
web front and end a cluster of databases. I'll put my thinking cap on.

Tim

Ross Scanlon

unread,
Oct 4, 2022, 12:11:33 AMOct 4
to osm-...@googlegroups.com
I've thought the same.

I have a second server in a totally different location (about 2000kms
apart) which could be used.  This is currently where I'm working at the
moment.

It currently has a mapnik instance on it and had been updating from the
minutely diffs.

It runs on Ubuntu 18.04, Postgres 9.5

It would be also nice to have failover on the dns so that fosm.org 
map.fosm.org  api.fosm.org  etc goes to the current server.

At least with failover on the dns I could put up a message that the api
is not available and the maps are static.  I only update the actual maps
about every six months there is no automatic generation with map requests.

Cheers

Ross

Tim Sheerman-Chase

unread,
Nov 7, 2022, 9:33:52 AMNov 7
to osm-...@googlegroups.com
Ross

Thanks for getting the server back on line! It was not too much trouble?

Tim

On 07/09/2022 01:11, Ross Scanlon wrote:
>

Ross Scanlon

unread,
Nov 7, 2022, 6:15:47 PMNov 7
to osm-...@googlegroups.com
Hi All,

Server was back online.

However one of the two ssd's that hold the operating system in a raid
array had failed.

Now the other one has failed.  So no server again.

They have been in continuous use for 5 years.

I do have a ssd here which I can replace them with but will have to
reinstall the operating system, hopefully today.

Fortunately the database and everything else is ok.  They are on a
separate raid arrays.

Cheers

Ross

Ross Scanlon

unread,
Nov 8, 2022, 4:19:39 PMNov 8
to osm-...@googlegroups.com
Hi All,

The api server is back online.  However if you were using https for
access in josm that's not back yet.  Just change to http to access it.

I'll get the rest of this sorted during the week.

Also maps are not available yet likewise parts of the website.

Cheers

Ross

Ross Scanlon

unread,
Nov 8, 2022, 10:02:25 PMNov 8
to osm-...@googlegroups.com
Hi All,

https is back now.

Cheers

Ross
Reply all
Reply to author
Forward
0 new messages