[Top][All Lists]

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [GNUnet-developers] downtime

From: Christian Grothoff
Subject: Re: [GNUnet-developers] downtime
Date: Sat, 13 Apr 2019 09:39:20 +0200
User-agent: Mozilla/5.0 (X11; Linux x86_64; rv:60.0) Gecko/20100101 Thunderbird/60.6.1

On 4/13/19 9:04 AM, address@hidden wrote:
> Christian Grothoff transcribed 2.4K bytes:
>> Dear all,
>> TUM has to move to another rack, so there will be some
>> planned outage tomorrow. Hopefully only for a few hours, but with an
>> uptime of 723 days I expect "fun" on reboot.
>> Best,
>> Christian
> Can you write another Email about what happened during the
> move or at least a public explanation why things look the way
> they do now? It will be less irritating for people who are not
> in the loop for this.

Well, to make a long story short, the /home directory was encrypted as
it was supposed to primarily contain personal data & inboxes when the
server was setup. Both encryption key copies were lost due to HW
failure, and no, it was dd'ed from /dev/random, so no chance of brute
forcing it. /home had no backups as it was (1) too big (see below on new
HDD) and (2) was presumed sufficiently redundant (RAID 1+1+1, 2 keys).

Additionally, when we moved from Subversion to Gitolite, the data that
used to be in /vcs or /var ended up in /home. Hence also the
Gitolite-managed Git repos were lost (note that the research data we
still kept in /vcs is unaffected, as that was the location for
non-private data by design...).  Finally, while the Drupal site database
is fine, /home included all uploads and the Drupal configuration, so is also down until someone finds the time and energy to
restore it.

Irnoy: The 8 TB replacement HDD for moving the data to a 2nd new server
in Bern (as I was quite uncomfortable with managing an 5 year old server
remotely with no physical access from the team for many years) had been
ordered a few days earlier and arrived _literally_ a few hours too late
to help. Anyway, the Bern site has right now IPv6 trouble, but once that
has been sorted out and the sysadmins have made progress, we should have
_two_ sites with copies of all of the important data. Oh, and thanks to
Andreas we're this time having three copies of the encryption key used
on the partially restored host.

Attachment: signature.asc
Description: OpenPGP digital signature

reply via email to

[Prev in Thread] Current Thread [Next in Thread]