FastMail Server issues

Aimlink · 22 May 2006, 09:20 PM

Last night I stopped reading mail because of connection issues with the server. Timeouts etc. Definitely FastMail since I get FastMail error pages and not generic timeouts.

Now, this morning it's the same thing.

MagicDavid · 22 May 2006, 09:26 PM

ARGH!!!! This is starting to wind me up now!

www.fastmail.fm - An error has occurred

Unfortunately an unrecoverable error occurred while trying to retrieve the requested page. We have logged this error and will try and fix the problem as soon as possible.

If you are really curious to know, the error was :

Failed to create screen: Mailbox, Could not connect to IMAP server. Please try again shortly. at /home/mod_perl/hm/ME/Mailbox.pm line 263.

davidbstanley · 22 May 2006, 09:27 PM

Same for me - No IMAP.

james374 · 22 May 2006, 09:28 PM

Quote:

Failed to create screen: Mailbox, Could not connect to IMAP server. Please try again shortly. at /home/mod_perl/hm/ME/Mailbox.pm line 263.

no sign of anything on the blog...when will it be back up again?

NJSS · 22 May 2006, 09:28 PM

It's working for me in the UK - BUT very slowly - again.

MagicDavid · 22 May 2006, 09:31 PM

"FastMail provides fast, professional and reliable email" - come on guys - this is the second time in the last few days!

MagicDavid · 22 May 2006, 09:41 PM

Back and up to full speed again

**ReuvenNY** · 22 May 2006, 10:13 PM

My service is ok, but in the past several days it has been slow. Everything, inluding opening and deleting emails, takes a few seconds more. I wonder if it's related.

JasonWard · 22 May 2006, 10:23 PM

Well the weekend has been and gone, Rob did say that had an additional server that was being mounted into the rack this weekend, so perhaps any hour now we will feel impact of its implementation and slow responses will go away...

Rob?

Jason

Rip Slagcheek · 23 May 2006, 12:38 PM

people must be lazy today. usually there are 4-5 pages posted within 30 minutes of an outage

NJSS · 23 May 2006, 02:26 PM

I think that the point probably is that this isn't a total outage, rather an outage affecting a percentage of users, and an invrease in response time affecting everyone?

It would be interesting and helpful to know how this is percieved by FM staff?

robmueller · 23 May 2006, 05:32 PM

It's being hit by the "the difference between theory and practice is always more in practice than the difference between theory and practice in theory"

We've got a plan to completely relayout our entire cabinets, which will make better use of space, allow us to completely mirror all servers, monitor the power usage on all circuits, have dual uplinks, etc. We've been working with NYI to get this done and they're saying this weekend or next weekend, so real soon.

In the meantime, we've been setting up some of this stuff already, such as the replicated frontend hearbeat servers. Now it turns out that as nice as it all sounds, and as well as the testing went, somehow production systems always throw more at you than you expect. heartbeat has been a bit of a cow, causing more problems currently than it's solved. Annoying. (I'll let Bron ***** about heartbeat in another post when he has time...)

Over the last 24 hours this has caused a couple of issues, including:

1. Both heartbeat servers being down at once and not accepting any connections
2. One server being down, and the other not taking over the IP, so half the users were down
3. Because of the "split" situation, half of the backends pings failed, and restarted themselves as well

So yes, there's definitely been a bit of up and down in the last 24 hours, but we've had everything mostly back up within 5-10 minutes if there's been a problem.

As always, we like to avoid downtime, it's painful and annoying, but we're also building what we think will be a great infrastructure to move forward with as well!

FYI we've also been working on the replicated backend servers, and after a bunch of testing we've moved about a hundred people on to those. That's been working well, we'll probably move some more users shortly...

Rob

**Gankaku** · 23 May 2006, 09:40 PM

I was in to take advantage of the special going on

And when I went to options I got:

"Oops, we can't open the web page you requested ...
- Error Unknown - Unknown error with URL"

I'll check back later.

JamesHenderson · 23 May 2006, 10:03 PM

still up and down (posting from England at 12:40 GMT).

tomkeifer · 23 May 2006, 11:26 PM

I have noticed that it is running a little slower then usual, but not to bad.

22 May 2006, 09:20 PM	#1
Aimlink Master of the @ Join Date: Oct 2005 Location: Here and Now... Posts: 1,078	FastMail Server issues Last night I stopped reading mail because of connection issues with the server. Timeouts etc. Definitely FastMail since I get FastMail error pages and not generic timeouts. Now, this morning it's the same thing.

22 May 2006, 09:26 PM	#2
MagicDavid Senior Member Join Date: Aug 2005 Location: England, UK Posts: 164	NOT AGAIN!!! ARGH!!!! This is starting to wind me up now! www.fastmail.fm - An error has occurred Unfortunately an unrecoverable error occurred while trying to retrieve the requested page. We have logged this error and will try and fix the problem as soon as possible. If you are really curious to know, the error was : Failed to create screen: Mailbox, Could not connect to IMAP server. Please try again shortly. at /home/mod_perl/hm/ME/Mailbox.pm line 263.

22 May 2006, 09:27 PM	#3
davidbstanley Cornerstone of the Community Join Date: Aug 2002 Location: Kent, UK Posts: 693	Same for me - No IMAP.

22 May 2006, 09:28 PM	#5
NJSS Master of the @ Join Date: Jul 2002 Location: Hampshire, UK Posts: 1,238	It's working for me in the UK - BUT very slowly - again.

22 May 2006, 09:31 PM	#6
MagicDavid Senior Member Join Date: Aug 2005 Location: England, UK Posts: 164	"FastMail provides fast, professional and reliable email" - come on guys - this is the second time in the last few days!

22 May 2006, 09:41 PM	#7
MagicDavid Senior Member Join Date: Aug 2005 Location: England, UK Posts: 164	Back and up to full speed again

22 May 2006, 10:13 PM	#8
ReuvenNY Moderator Join Date: Mar 2002 Location: New York Posts: 4,259	My service is ok, but in the past several days it has been slow. Everything, inluding opening and deleting emails, takes a few seconds more. I wonder if it's related.

22 May 2006, 10:23 PM	#9
JasonWard Cornerstone of the Community Join Date: Mar 2004 Location: London, UK Posts: 834	Well the weekend has been and gone, Rob did say that had an additional server that was being mounted into the rack this weekend, so perhaps any hour now we will feel impact of its implementation and slow responses will go away... Rob? Jason

23 May 2006, 12:38 PM	#10
Rip Slagcheek Member Join Date: Jan 2006 Location: NYC // Texas Posts: 72	people must be lazy today. usually there are 4-5 pages posted within 30 minutes of an outage

23 May 2006, 02:26 PM	#11
NJSS Master of the @ Join Date: Jul 2002 Location: Hampshire, UK Posts: 1,238	I think that the point probably is that this isn't a total outage, rather an outage affecting a percentage of users, and an invrease in response time affecting everyone? It would be interesting and helpful to know how this is percieved by FM staff?

23 May 2006, 05:32 PM	#12
robmueller Intergalactic Postmaster Join Date: Oct 2001 Location: Melbourne, Australia Posts: 6,102 Representative of: Fastmail.FM	It's being hit by the "the difference between theory and practice is always more in practice than the difference between theory and practice in theory" We've got a plan to completely relayout our entire cabinets, which will make better use of space, allow us to completely mirror all servers, monitor the power usage on all circuits, have dual uplinks, etc. We've been working with NYI to get this done and they're saying this weekend or next weekend, so real soon. In the meantime, we've been setting up some of this stuff already, such as the replicated frontend hearbeat servers. Now it turns out that as nice as it all sounds, and as well as the testing went, somehow production systems always throw more at you than you expect. heartbeat has been a bit of a cow, causing more problems currently than it's solved. Annoying. (I'll let Bron ***** about heartbeat in another post when he has time...) Over the last 24 hours this has caused a couple of issues, including: 1. Both heartbeat servers being down at once and not accepting any connections 2. One server being down, and the other not taking over the IP, so half the users were down 3. Because of the "split" situation, half of the backends pings failed, and restarted themselves as well So yes, there's definitely been a bit of up and down in the last 24 hours, but we've had everything mostly back up within 5-10 minutes if there's been a problem. As always, we like to avoid downtime, it's painful and annoying, but we're also building what we think will be a great infrastructure to move forward with as well! FYI we've also been working on the replicated backend servers, and after a bunch of testing we've moved about a hundred people on to those. That's been working well, we'll probably move some more users shortly... Rob

23 May 2006, 09:40 PM	#13
Gankaku Moderator Join Date: Mar 2002 Location: Virginia, USA Posts: 3,265	I was in to take advantage of the special going on And when I went to options I got: "Oops, we can't open the web page you requested ... - Error Unknown - Unknown error with URL" I'll check back later.

23 May 2006, 10:03 PM	#14
JamesHenderson Cornerstone of the Community Join Date: Jan 2003 Location: Oxfordshire, UK Posts: 603	still up and down (posting from England at 12:40 GMT).

23 May 2006, 11:26 PM	#15
tomkeifer Member Join Date: Mar 2006 Posts: 79	I have noticed that it is running a little slower then usual, but not to bad.