Hello again.
At Sunday 31 January 2010 14:14:13 DaB. wrote:
> I can give no link to a jira-ticket, because jira is down since yesterday.
after jira is back now (thanks to River), the ticket-url is
https://jira.toolserver.org/browse/MNT-205. I will also reboot cassini for the
same case too, so I added my original mail below for the OSM-folks.
Sincerly,
DaB.
>Hello all,
>
>for an important kernel-update, nightshade will reboot around 23 o'clock UTC.
>The downtime should be short, but all running programms on nightshade will of
>corse stop. So please make sure that they restart after the reboot or move
>them temporary to willow (caution. solaris! ;)).
>
>I can give no link to a jira-ticket, because jira is down since yesterday. I
>will send a email after the reboot to inform you all.
>
>Sincerly,
>DaB.
>
--
wp-blog.de
Hello all,
for an important kernel-update, nightshade will reboot around 23 o'clock UTC.
The downtime should be short, but all running programms on nightshade will of
corse stop. So please make sure that they restart after the reboot or move
them temporary to willow (caution. solaris! ;)).
I can give no link to a jira-ticket, because jira is down since yesterday. I
will send a email after the reboot to inform you all.
Sincerly,
DaB.
--
wp-blog.de
Hi,
Over the next couple of weeks we will be making some changes to the way that
wiki databases work on the Toolserver.
Currently, each cluster (s1..s6) has only a single Toolserver slave. (For
example, s1 is on yarrow and no other server). We will be adding an
additional server to each cluster, so each cluster has two servers. The
existing server will be designated the 'standard' server, and the new server
will be a 'fast' server.
The standard server can be used as it is now, with no changes. The 'fast'
server will be the same as the standard server, with two exceptions:
* creating user databases will not be allowed
* queries running longer than 60 seconds will be automatically killed
The intent here is that by preventing long running queries on the fast
server, replication lag will be much less likely, and queries that need the
most up-to-date data, which are usually also fast queries, will be able to
use the fast server.
If either server fails, its queries will be redirected to the other server.
So, if the s1 fast server fails, all fast queries will go to the standard
server. If the standard server fails, all queries will go to the fast
server.
If you want your queries to run on the standard server, you do not need to
do anything. The current server for each cluster, where the existing
'sql-sX' aliases point, will become the standard server.
if you want your queries to run on the fast server, you should connect to
'sql-sX-fast' instead of 'sql-sX'. You cannot connect to the fast server
using DNS aliases; you must use the sql-sX aliases.
Currently, no fast servers are available, so all sql-sX-fast aliases point
to the standard server. The first cluster to get a fast server will be s1,
which I hope will happen next week. One cluster will not get its fast
server until all projects have finished migrating from the stable server, as
the stable server host will be converted into the 3rd fast server.
- river.
Hi,
Our trained monkeys have been working non-stop for the last week on a new
piece of software for the Toolserver: the jobserver. This software is now
ready for testing.
The jobserver is an alternative to cron/phoenix for starting long-running
jobs. Instead of having to invoke phoenix regularly in cron, you can
instead add your job to the jobserver, and it will be (re)started
automatically, even if the server crashes or reboots.
This is still very much pre-alpha, but I've installed the jobserver on
willow for testing. You can read more about the jobserver and how to use it
on the wiki:
<https://wiki.toolserver.org/view/Jobserver>
Please report any problems on the wiki page. While you should feel free to
use and test the jobserver as much as you want, do remember that it may
still have serious problems that result in your jobs not running when you
expect, or maybe worse.
In the future the jobserver will have additional functionality, including
scheduled jobs (to replace cron with a single integrated solution). If you
have any other feature requests, you should list them on the wiki.
- river.
Hi,
Tonight I will upgrade the web server software on wolfsbane. Total downtime
for the upgrade should be around 5 minutes. There will be no user-visible
changes.
This issue is being tracked in JIRA as MNT-177.
- river.
Hello all,
on saturday night, I will install all non-critical updates of debian-packages
on nightshade. This time, twisted will be updated too and it will have a major
release jump from 8 to 9.
If you use twisted, please read the changelog (I attached a copy to the
maintaince-ticket) and check if there could be problems. If you found a
problem, please leave me a notice and I will delay the update until you can
fix the problem with your code.
The update will also affect some other packages (for details, see maintaince-
ticket), but there should be no problems AFAIS.
The maintaince-ticket is at https://jira.toolserver.org/browse/MNT-174.
Sincerly,
DaB.
--
wp-blog.de
Hi,
The above two wikis will be dropped from the cluster, as they are obsolete.
They have already been removed from the toolserver.wiki table.
- river.
Hi,
We will be performing maintenance on MySQL tonight to fix the 'host is
blocked because of too many connection errors' problem. This is being done
earlier than expected because it's required for the new load-balanced MySQL
servers. The total downtime should be under 20 minutes per server.
This issue is being tracked in JIRA as MNT-165.
- river.
Hi,
I am about to decommission cache.stable.toolserver.org. Requests to the
host will still work as before, but will go straight to the web server
without being cached.
This may return in the future as a cache for toolserver.org instead.
- river.