Global DNS outage

Hi all,

it appears that the machine hosting the forum and nameserver for became inaccessible some time between sunday, 6th and monday 7th. We tried to contact the admin about the problem but didn't receive a direct response so far. According to a statement made by Zoltan, a hardware fault is the probable cause for the outage.

Since the server didn't come back online within the last ten hours and since we didn't hear about any ETA on repairs yet, we decided to take the matter to the SPI Inc., the non profit organization handling the domain, and asked them to redelegate the zone to the Digital Ocean name servers which are also already hosting the domain.

At this point in time we unfortunately cannot give any estimation on when these issues will be sorted out as we're still waiting for a positive confirmation from SPI. Depending on the reaction time of SPI's registrar and the time it takes for the DNS changes to propagate it could happen anytime within the next 12 to 96 hours.

It is also possible that the server hosting the authoritative DNS is getting online sooner, intermittently fixing DNS resolving.

Update Tuesday, May 8th 07:00 CEST

The authoritative name servers have been changed two hours ago and should slowly propagate down the DNS chain now.


If you find yourself to be unable to reach any of the hosts, e.g. or and happen to run an OpenWrt or LEDE router, you can temporarily configure it to use the DigitalOcean DNS servers to resolve * addresses:

  • Log into LuCI
  • Navigate to Network --> DHCP and DNS
  • Enter the following value into the DNS forwardings field: /
  • Press Save & Apply


1 Like

If you prefer the CLI approach, use the following commands to forward *

uci add_list dhcp.@dnsmasq[0].server=/
uci commit dhcp
/etc/init.d/dnsmasq restart

If you happen to run dnsmasq on non-OpenWrt systems or if you prefer to use the native configuration, you can add server=/ to /etc/dnsmasq.conf as well.

1 Like

the sad part is;
i actually prefer the vision of a pitiable admin at spi having a really shitty monday over attributing this to malice or negligence

You know you can speed up the process considerably if you still have access to the previously delegated slvaes, and can reconfigure them to pull the zone from the new master? Even only one of them would help at this point, when all the other servers are "gone" (master down and slaves failing). It would bring back online immediately without the need to wait for delegation updates and propgation delays

I made a post here:

I now realize the REAL master is offline.

This will not work, unless the real NS records are changed.

The domain info at the registrar has finally been updated a few hours ago (2018-05-08T03:03:51Z) with new name servers.

The correct new info (name servers set for

Domain Name: OPENWRT.ORG
Domain ID: D104186352-LROR
Registrar WHOIS Server:
Registrar URL:
Updated Date: 2018-05-08T03:03:51Z
Creation Date: 2004-04-02T22:32:12Z
Registrar Registration Expiration Date: 2019-04-02T22:32:12Z
Registrar: TUCOWS, INC.

Things should get fixed soon (except regarding that is apparently on the same physical hosts as the old name servers at

1 Like

I still can't update feeds from Is there an IPV4 address to use instead of a domain name?