What's happening with the Forums

ZTD

jamesd wrote: »

NFS security (and MySQL security) in a colo. Hmm, I wonder just how good this colo is at keeping customers apart so one can't attack the others, perhaps as a result of a compromised other customer?

Well that depends really. It is possible to put a customer on their own VLan and firewall all the traffic onto/off of that VLan using something like a netscreen/juniper. But not many bother to do that. If this site is a virtual host on VMWare, then probably a great deal less than that.

jamesd wrote: »

I hope they at least don't run MySQL on the standard port (known target, easy protection), do use the latest more secure client protocols and do follow the basic account security recommendations that MySQL makes.

Changing the port isn't going to help much. If they've firewalled it off from "out there", then the only thing that is going to be able to access it is a compromised host. That usually tends to mean there is a human brain in charge, who at the worst can always read the co-lo's documentation as to which port they're running MySQL on...

jamesd wrote: »

MySQL will buffer writes for a while in disk full cases, no bets on how NFS might break to break that.

Writes to a full disk return immediately with an error code. Writes to a "gone-away" NFS mount block until it comes back (with exceptions, but I'll ignore them ) So you can end up with all your writer threads "stuck".

jamesd wrote: »

But it's a trade-off: do you want an outage that's going to be really obvious or potential loss of recent data?

It's not always that clear cut. If you're using the database to log events, any server outage *is* losing recent data.

jamesd wrote: »

Depends; I tend to prefer obvious trouble so it's noticed and fixed quickly. But since you can use replication to make it irrelevant that's the better answer.

Haven't looked at the finer details of semi-sync replication but it probably assumes a reliable network, or at least that one of them is reliable - say a dedicated connection via the second network port that's common on servers.

Anything that relies on reliability is unreliable...

jamesd wrote: »

I've never used MS NFS. You make me happy that I haven't.

I was going to say "it's not *that* bad" - but it is. It's dire. I think "NFS daemon for the Vic 20" probably works better...

littlejaffa

so um...who is this wonderfull server company then?
just so i can avoid them of course!

Hope your asking for a refund!! not fit for purpose should do it!

Potsdamerplatz

Thanks to Martin and all the team and posters who make this such an important website.

Shame on the hackers for ruining it for everyone - especially as many of the people who ask for help on this forum are vulnerable, desperate and have nowhere else to turn to for impartial advice and information.

Let the rebuilding work begin.

As Winston Churchill said many years ago - we will never surrender or give in to tyranny.

:beer:

Tiff_2

Hi Martin,

and !



Much Love,
Tiff xxx

Former_MSE_Webmaster

Hi folks,

Tonight we did some work trying to restore the old posts. We think've succeeded to a reasonable extent although there may still be a some missing, as Martin explained in the OP.

We've repaired some other portions of the Forum as well, including the last post links and the page counts on long threads. If you spot any errors, please post them here.

Hopefully, we're almost back to where we were.


Webby

Alfie_E

Well… that’s some improvement! Previously, as redux reported, we were missing all the posts between those numbered 77931 and 5595740. Now, the gap starts just after post number 5587385. The loss seems to have been reduced to only some eight thousand that were mostly posted on June the 28th. (Not every number has a post associated with it, as people change their minds and delete their own posts and spammers’ posts get wiped by the abuse team, but most do.)

That’s a 99.85% improvement!!!

Hurray for Webby! Hurray for back-up tapes! Hurray for Pocoyo!

fermi

:T:T:beer:

Well done.:D

There's lots of stuff to check out, but it's looking very good so far.

Quite a lot of stickies and important first posts will need to be updated again to the cached versions we have recovered from Google and elsewhere, but that was expected.

Excellent Job.

Edinburghlass_2

Does this mean no more [STRIKE]dirty[/STRIKE] techie talk from jamesd and ZTD :rotfl:

peter999

Alfie_E wrote: »

Well… that’s some improvement! Previously, as redux reported, we were missing all the posts between those numbered 77931 and 5595740. Now, the gap starts just after post number 5587385. The loss seems to have been reduced to only some eight thousand that were mostly posted on June the 28th. (Not every number has a post associated with it, as people change their minds and delete their own posts and spammers’ posts get wiped by the abuse team, but most do.)

That’s a 99.85% improvement!!!

Hurray for Webby! Hurray for back-up tapes! Hurray for Pocoyo!

post #5587385 = 28-06-2007, 3:47 AM
post #5595740 = 29-06-2007, 1:02 AM

5595740 - 5587385 =8355 posts missing ??

peter999

redux

redux wrote: »

The team who will sort this out will rely on good comprehension and inter-personal skills, and a calm approach without hasty judgements.

... and mastery of understatement about their efforts, it appears:

MSE_Webmaster wrote: »

We think've succeeded to a reasonable extent although there may still be a some missing, as Martin explained in the OP.

:T