Darknetlive

  1. Home
  2. Posts
  3. Dread is Migrating to a New Host

Dread is Migrating to a New Host

Dread administrators are actively migrating Dread to a new host and the downtime is “nothing to be be concerned about,” according to a message from one of the admins. An issue with the current host unexpectedly brought Dread down and the host provided “less than adequate responses” when asked about the incident.

Edited to add an update from HugBunter.

Paris wrote, in a message on Envoy (here) and on Dark.Fail (here):

So we got some host problems. Nothing to be be concerned about but we can’t stay with the host the current cluster is on. In one swoop the whole cluster went offline for really no reason. Reaching out and finding out why yielded… less than adequate responses to my tastes.

Simply put I don’t trust them anymore and until I can get the distributed cluster design to work we are stuck with this kind of localized bullshit. Hopefully the next host will be not as much as an asshole as this one is. We are probably going to lose about 4 days of content due to this (unless hug has a more recent full backup). Dread is a free service supported by some good advertisements and of course all ads will be extended ontop of the downtime.

Come and say it with me “DISTRIBUTED HOSTING IS THE FUTURE BECAUSE HOSTS CAN BECOME ASSHOLES”

Cheers,

Paris

So we got some host problems...

So we got some host problems...

In response to a question about the data loss, Paris wrote:

Basically it will be like a time reversal. The site will just go back to the snapshot date when the backup was taken place. So if you created your account made a post or comment in the last four days that information will be lost into the abyss. Shit happens. At least it’s not two weeks this time.

HugBunter posted a message on the DreadAlert subreddit on January 10.

We have experienced a server crash which has affected our whole cluster and we expect to have an extended period of downtime right now, spanning into the weekend.

Currently investigating as to the cause and any damage we may experience due to this. Hopefully there will be no data loss, however it appears there may. We have recent backups which will keep any data loss at a max of 3-4 days. Of course this would still not be a blow that this will soften any further due to the extremely increased user activity Dread has had in recent weeks
due to perfect uptime and platform speed increases. Any accounts, posts, comments from the last few days may be lost.

Sorry for the inconvenience caused, I’ll be pushing a planned feature update once service is restored.

Show Paris' Signed Message
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA256

So we got some host problems. Nothing to be be concerned about but we can't stay with the host the current cluster is on. In one swoop the whole cluster went offline for really no reason. Reaching out and finding out why yielded... less than adequate responses to my tastes.

Simply put I don't trust them anymore and until I can get the distributed cluster design to work we are stuck with this kind of localized bullshit. Hopefully the next host will be not as much as an asshole as this one is. We are probably going to lose about 4 days of content due to this (unless hug has a more recent full backup). Dread is a free service supported by some good advertisements and of course all ads will be extended ontop of the downtime.

Come and say it with me "DISTRIBUTED HOSTING IS THE FUTURE BECAUSE HOSTS CAN BECOME ASSHOLES"

Cheers,
Paris
-----BEGIN PGP SIGNATURE-----
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=ptYo
-----END PGP SIGNATURE-----

Show HugBunter's Signed Message

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA512

We have experienced a server crash which has affected
our whole cluster and we expect to have an extended
period of downtime right now, spanning into the
weekend.

Currently investigating as to the cause and any damage
we may experience due to this. Hopefully there will be
no data loss, however it appears there may. We have
recent backups which will keep any data loss at a max of
3-4 days. Of course this would still not be a blow that
this will soften any further due to the extremely
increased user activity Dread has had in recent weeks
due to perfect uptime and platform speed increases.
Any accounts, posts, comments from the last few days
may be lost.

Sorry for the inconvenience caused, I'll be pushing
a planned feature update once service is restored.
-----BEGIN PGP SIGNATURE-----
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=rEdX
-----END PGP SIGNATURE-----

And in .txt format here and here.