Mail issues

[scook94]

Just out of curiosity, how many hundreds of gigs is the HD on this server, that takes 2 days to run a diskcheck?

Would be surprised if it wasn't terabytes...

[Malc]

Not sure what the capacity is, Malc, but at 16:51 this afternoon, it had apparently checked 67.3M files.

Thats hell of a lot of files

[Simon]

There's a hell of a lot of emails queueing!

[Lance]

Indeed. And when Trevor does come back, they'll release the emails in a managed quantity to make sure he doesn't get overloaded. Will probably take half a day at least for them all to be put through.

[Sebby]

The last thing we want is for them to all be released and crash the server.

[D-Dan]

I hate this thread and I hate this forum. As a result, I just committed myself to £4.99 per month hosting

If I hadn't been here reading, the idea wouldn't have been in my head.

So a special thanks to Rik for costing me yet more dough. 

Steve

Edit: I just worked it out. That's nearly three drops of petrol 

[Sebby]

Maybe Rik will pay for it.

[Simon]

Rill is a well renowned expert at spending other people's money.  You dropped your guard there, Steve. 

[john]

... also, it seems, rather lucky, Zimmer!   

Not sure what the capacity is, Malc, but at 16:51 this afternoon, it had apparently checked 67.3M files.

So that's all Rik's posts done then, hopefully they're now doing ours   

[sebt]

I spoke to James at Idnet today; apparently a number of processes on the attacked server had gone rogue so they chose to reboot it. During reboot the server proceeded into an automatic file check (fsck) which would suggest that the filesystem(s) had been corrupted and/or the journals marked unclean. Either way, from my experience (I work extensively with Linux in addition to various Unix flavours from my past) an fsck operation's time to complete depends on a number of factors: The size and complexity of the filesystem (how many files, folders, etc.) and the filesystem type (typically in Linux this would be ext2/3, XFS or ReiserFS). An fsck taking over 24 hours and counting might suggest a very large ext3 filesystem with millions of small files and directories that has not been metadata cleaned for a long time, with possibly heavy metadata corruption.

Unfortunately, they have no real idea how long this takes and until complete will not know the extent (if any) of the damage. The good news being that apparently there are backups taken every other night, and that incoming email to the (downed) server is buffered by the inbound SMTP gateway, so it will not ping back but just wait for the offending server to come up before delivering all of the buffered mail.

In the meantime, users who want to send outgoing email can use smtp.idnet.net or smtp.idnet.co.uk as the SMTP gateway.

Perheps I should head up to London and assist, though it's rather a long way from cornwall

Seb

[Simon]

Thanks, Seb, and I note you haven't received a welcome Karma, so here it is. 

[Sebby]

And one from me too...

[Lance]

I know there is certainly a lot of small files - as our Simon had said, by yesterday the server had checked 63.3 million files!

[Rik]

Part of the problem, I gather, is the increasing use of IMP, and users not subsequently clearing their mailboxes out. A 100MB can hold a lot of files.

[Rik]

So a special thanks to Rik for costing me yet more dough. 

NP, Steve - it's what I do best.

[Gary]

So still no ETA Rik and still a backlog of email building up?

[Rik]

I've heard nothing since yesterday, Gary, I'll be chasing soon.

[Steve]

Part of the problem, I gather, is the increasing use of IMP,

Is that the same as IMAP?

[Rik]

It's exactly the same as IMAP when I don't type the A, Steve. 

[Steve]

Sorry, this an imp http://upload.wikimedia.org/wikipedia/commons/e/eb/TheLincolnImp.jpg

[Rik]

I see you've been gargling, Steve.

[Gary]

I've heard nothing since yesterday, Gary, I'll be chasing soon.

Cheers Rik, I have been looking at hosting services and seen pages of horror stories about 1&1 I know you have no issue but with them, but knowing my luck   and after the present outage im a little apprehensive, in the past I would have gone for IDNet hosting, but having got it in the ear from a few people i told to use idnet, a: I feel petty awkward now, causing b: my trust to be knocked with them. Hopefully things will move on today as I know they do care and want things up and running for their clients. Seems to me like Trevor needed some maintenance anyway maybe

[Rik]

There will have to be an inquest to find out what went wrong with Trevor, Gary, if it ever gets through the self-check.

I would say you get what you pay for with 1&1, but I've had 17 months without a problem paying less than IDNet charge (ie 5 boxes for £0.69pm). Take a look at Purple Cloud, ISTR that Ann uses them.

[Gary]

There will have to be an inquest to find out what went wrong with Trevor, Gary, if it ever gets through the self-check.

I would say you get what you pay for with 1&1, but I've had 17 months without a problem paying less than IDNet charge (ie 5 boxes for £0.69pm). Take a look at Purple Cloud, ISTR that Ann uses them.

Same as most things in life Rik,  cheers for the advice

[Steve]

Can you image the average computer user waiting this long for a disk check to complete,I would be sorely tempted to press the power button and say b***** it