Hi,
There is currently an open issue that is preventing updates to the
Toolserver email configuration from propagating to the mail server. This
means that email forwarding for newly created accounts is not working, and
changes to your Toolserver email address (using setmail) will not take
effect. The issue is somewhat complicated and I will be away most of this
week, so this won't be fixed until next week at the earliest. Sorry for the
inconvenience.
This issue is being tracked in JIRA as MNT-65.
- river.
Hi,
You can ignore this message unless you have C or C++ programs or libraries
linked directly against 'libjpeg', the IJG JPEG library. (The vast majority
of users will not have any such programs.)
libjpeg on the Solaris systems (willow and wolfsbane) has been upgraded from
version 6b to version 7. This means the old libjpeg.so.62 no longer exists;
it is replaced with the incompatible libjpeg.so.7. You will need to
recompile and relink your programs to use the new libjpeg.
This doesn't affect programs that use libjpeg indirectly, e.g. via the
ImageMagick C/C++ interfaces, since all TS-provided software depending on
libjpeg has already been relinked.
- river.
Hello all,
yesterday we noticed, that our s1-cluster (that's enwp) is corrupt (that means
that we don't have all data and/or not in the right way as the master).
That means that we will need a re-important soon. Please look at
https://jira.toolserver.org/browse/MNT-104
to see the process.
Sincerly,
DaB.
P.S: River told me now, that we can perhaps minimize the downtime of s1 this
time - we will see.
--
wp-blog.de
Hi,
PHP was upgrade to 5.3.1. At the same time, the PHP PDO MySQL module was
enabled. This should have no impact on users (unless you want to use PDO).
- river.
Hi,
MySQL on yarrow (s3, s4, s6) crashed and restarted. This is a known problem
and has happened before; however, we've now made a configuration change that
might prevent it from happening in the future.
This issue is being tracked in JIRA as MNT-58.
- river.
Hi,
The web server wolfsbane will be rebooted later tonight to change a
configuration setting. Web tools will be offline for about 10 minutes while
this is in progress.
This issue is being tracked in JIRA as MNT-90.
- river.
Hi,
Later tonight (UTC) I will upgrade the Toolserver wiki to the current
MediaWiki trunk version. There will be a brief outage of the wiki while
this is done.
This issue is being tracked in JIRA as MNT-78.
- river.
Hi,
As announced two months ago, we are planning to EOL the stable server in
favour of multi-maintainer projects on the normal Toolserver. The following
projects are still running on the stable server:
* geohack
* wma
* delinker
We intend to repurpose the hardware currently used for the stable server to
provide redundancy database replication, but this can't happen until all
projects have migrated off the stable server.
- river.
Hello all,
today we run into a deadlock-situation on rosemary (that's the sql-server for
s1 (that's primary the english wikipedia)) where many queries seems to block
each other (see [MNT-60] for some details).
To end this deadlock, I killed all queries, even the *SLOW-OK* ones. So if you
had run a query between 12 o'clock and 16'o clock and it was killed, that was
me, please restart it if needed.
Sincerly,
DaB.
--
wp-blog.de
hi,
we had an outage of ~5 hours affecting all login servers from 04:15 UTC
to around 09:30 UTC. the issue was fixed fairly quickly once noticed,
but it took some time for an admin to be contacted.
the outage was caused by a misconfiguration on the cluster which
prevented switching the NFS service from a failed node to a working one.
the misconfiguration has been fixed, so failover should work correctly
in the future. (this was our first unscheduled cluster failover, so it
hadn't been tested before.)
a full description of the outage is available in MNT-56.
- river.