I continue to be deafened by the silence from Plusnet.

You have to remember this is not an official channel of support.  We often get help if we're nice to people!!

But the official support channel does not even seem to understand how Postini Quarantine actually works judging from previous replies to messages I have sent it or phone calls I have made to it.

Are you saying I am using an unsupported service and should use the Plusnet IMAP spam folder option instead if I want a supported spam filtering service?

Being meek and nice isn't always the most effective method of getting results.  If the several people who wanted to get off the Spanair plane at Madrid after the first aborted takeoff last week had kicked up one hell of a fuss (instead of just asking politely as several did) to the point of creating a major disruption I am sure they would have achieved their goal of being removed from the plane and thereby saving their lives in many cases.

Not at all.  Postini quarantine is an option in MMM and, as such, I'm sure it's supported in the same way as the rest of the email service.  But the forum is a community help and gossip channel.  To access support reliably (not necessarily the same as accessing reliable support!), you should raise a ticket.

Also I'm saying ... sorry I can't help with your queries....

Chris

Firstly, apologies for the lack of input on this thread of late. As Kelly has mentioned elsewhere, the Comms Team have been really busy over recent weeks helping out the CSC guys. This means we've had less time to contribute to the forums.

I've had a quick look through the last three pages or so of this thread. Hopefully I've covered everything below...


I think the 'essential functionality' aspect of this could be argued by Postini. 'Catch-all' quarantine accounts are considered by them to be a legacy feature. It's partly for this reason that we were never able to send the daily quarantine summary emails that we originally promised. I too would have liked to have seen this particular issue resolved quicker. I can tell you that the speed and quality of support we're provided by third parties will very much be at the forefront of any future partnership decisions we make regarding the mail platform.


There is a control panel, however this is not accessible on a per-user basis. This was a decision that we consciously made when the platform was originally implemented as it would have made it very easy to 'break' your spam settings.

If you want to change your password then enter it incorrectly. This gives you a forgotten password link. Click this and you will be sent a temporary password. You can then log into Quarantine and change this to a new password of your choice.


TBH, 'last 3 days' kind of makes sense to me. Presumably you'd be checking the Quarantine at least that frequently so it helps move the clutter from view. Are you sure that 'most end users'  regularly purge the contents of their Quarantine? Are you basing that on fact or assumption? Is it really that difficult to use the drop-down to filter based on broader criteria? I can see sense in some of your reasoning but I'm not sure it's got sufficient weight to actually spur Postini into action I'm afraid.


Just for the record, changing the default view isn't something you can do even with access to the end user control panel I referred to above. Is it not easier for you to filter using the 'Date' and 'To' columns? That's what I'd do. In fact different people are going to have different preferences. If we changed the default, guaranteed there'd be others out there who wanted it changing back!


Being able to select the default view on a per-user level would be ideal but it would be a request for additional functionality that we'd have to submit to Postini. AFAIK, there's nothing in the existing build that lets us configure this on either an individual or global basis.


I'm guessing it is cookie/session based. I have just logged in and out a few times using Firefox 3 and my preferences were saved. When I then tried using IE 7 the settings reverted back to their defaults.


I know nothing of this particular issue and I've been unable to replicate it on my account. Do you know if it's still happening?


There's no API to do this and as I mentioned previously, Postini consider 'catch-all' to be a legacy feature so building functionality around this isn't going to be high on their list of priorities I'm afraid.


Anybody else seeing this. My updates seem to be working OK from what I can tell, but I don't rely too heavily on the email notifications so I'm less likely to notice things like this.


The rate of false positives/negatives that we get from Postini has always been subject to debate and most will agree that it's something that could be improved on. I'd also be lying if I said that we hadn't been looking at alternative solutions. In fact there are a number of viable solutions that we've been trialling from before the introduction of Postini. I'm sure we'll here more about these within time.


spam@spamtraining.plus.com and notspam@spamtraining.plus.com respectively. A selection of the mail in these mailboxes is forwarded to Postini's own spam training addresses so it doesn't matter whether you use Quarantine or not.


Are these emails destined for mailboxes that actually exist? they've not been sent to a catch-all user that you've not got set up at the Plusnet side have they?


Ah, that clears that one up then so feel free to ignore my earlier assumption. I guess this could have changed with the roll-out of Postini's latest build.


Really? By definition I'd have thought that's exactly what it means. Under what circumstances would you want the messages delivered?


I think changes made on our systems are pushed to Postini immediately but there may be some delay at their side. I'd need to get clarification on this though as I'm not 100% sure.

To summarise, I know Quarantine isn't perfect. Please remember though that for all intents and purposes it's a proprietary platform. Making small changes isn't as easy as it could be as we don't own the code.

Some of the other solutions we've been testing allow us to contribute much more closely in the development cycle of future builds. Unless we adopt one of these approaches though, I think that we're always going to be left with the odd annoyance here and there.

Actually if you change the password of your main Plusnet mailbox login turns out it eventually rolls out automatically to Postini Quarantine two or three weeks later as I have now found has happened to my Postini Quarantine.


No I typically check it every two weeks on average because after whitelisting the non spam stuff Postini wrongly considered Spam (mainly bulk emailers like Ebay who do still let you unsubscribe if you want to so should not be considered spam) .  So you assume wrong about my own habits in respect of Quarantine use.  As you can't send me a summary email from Quarantine as Plusnet promised would happen when and how often I check Quarantine is bound to be erratic.


I'm basing it on normal deductive logical brainpower rather than customer service person think of "lets bat what the customer wants straight back in their face"

Of course most users will delete all the spam in the Quarantine folder each time they visit as otherwise it will be there again next time and they will have far more to check next time.


That is the summary of your whole reply.  We do it the way Postini wants to and not the way our customers want it to be and there are no user configurable options other than the white and black listed email addresses set through the Plusnet mail portal.


How would I identify spammers sorting by the Date and To columns?  spammers send on any old date and the messages are nearly always To some variant of my email address (mainly actual mailboxes with a minority of Catch All To addresses)  If I sort by Subject I immediately come across the same message loads of times with the same Subject from numerous different senders.  Ditto with the From address I easily spot the occasional genuine sender who needs whitelisting.  How would I do that if I sorted by the date or To columns (especially when nearly all the To is to me).

Are you sure you aren't just looking for typical smartass slam it back in my face customer service person method of response to try to ignore a perfectly legitimate request?


Why do you have to guess? Surely Postini gave you a detailed spec of how their interface works?


So we are forced to use Plusnet's own IMAP spam filtering box to stop Catch All messages getting through even though its less reliable than Postini by quite a long way.  And even though its Plusnet's fault that most Plusnet email addresses were stolen and targeted for spam.


Why change and start all over again.  Why not improve a service that does a lot already but just needs a few rough edges tidied up.


So what is going to be done to correct this error where I cannot deliver emails I want to deliver that are from blocked addresses.


Under the circumstances that someone who mainly sends junk circulars and also far too many of them then sends the odd useful message.


To summarise you have just ignored all my suggestions for improvement that would make the service better and said its really too much trouble and we don't care about the suggestions of our mere end users.

That I didn't know and I shall try and find out why it is. Off the top of my head there aren't any scripts that run that infrequently but I've no reason to doubt what you're saying.


As mentioned, there's system limitations as to why we can't send the summary emails. We could do this if we made some changes but that wouldn't involve us sticking with Postini.


And that may well be why we've never excluded the other alternatives. I've already mentioned the false positive/negative thing and this is another factor that plays a large part in our trialling of other solutions. Believe me, a lot of your frustrations are mimicked in house. I really don't like to point the finger but there comes a time when you have to weigh up the alternatives!


I guess that depends on your circumstances. Me personally, I'd sort by 'To' address first to can off the stuff that is addressed to non-mailboxes and then go by date. It's all down to user preference though really.


I'm sorry, but the last thing I tried to do was ignore your requests. On the contrary, I've tried to address each one as best I can! It may not involve the answers you were looking for but I've tried my best to reply in an honest and balanced manner none-the-less.


Not to that degree, no.


Which sums up why we haven't stopped working on this. No we don't expect you to rely on our in house solution, and yes we are looking at making this better.


We're committed to making things better, that's for sure. Having said that, if there's more suitable solutions out there then we'll go for them! There's at least a handful that have been integrated from before the advent of Postini.


If I'm honest this confuses me. I make the decision to receive email from a company or not. I wouldn't expect an anti-spam solution to selectively know which of said emails I wanted to receive. I might be missing the point, but it's kind of like a 'subscribe or don't subscribe' situation in my opinion.


That's really not the impression I intended on portraying. I've tried my best to address all the points you've made. We *are* listening to feedback, and there's a lot of discussion as to how we can make things better. There's only so much that Postini can do though...

I don't expect the anti spam solution to know which mails from a certain sender that is in the list of blocked addresses it should deliver I expect to do that manually myself using the Deliver button against the individual emails in Quarantine that I still want to receive from that sender.  But I can't because the Deliver button no longer delivers any quarantined emails that are from one of the Blocked Addresses that causes those emails to initially be consigned to spam despite not being trapped by Postini's own spam filtering algorithms.

By the way apologies for the possibly slightly intemperate tone of my earlier responses as I can see you have been very patient in going through them all.  Perhaps it had been a bad day and the news that Postini and Plusnet collectively could not respond at present to any of my issues on the spam filtering service was rather disappointing........

The message centre password gets set to your account password when you make changes to your Postini configuration, e.g. make changes to your black/whilte lists. I'm guessing it was two or three weeks after you'd changed your account password when you next made a change.

I can confirm that the Show All and number per page is controlled by a cookie. Mine remembers the last setting I made. I deleted the Postini cookies and it went back to the defaults.

I think you have a very strange interpretation of a "blacklist". To 99% of the populatiion, it means "I don't want to receive anything from this address - ever" so there is no way you'd want it delivered. The only reason it appears in the Message Centre is so that you can check that nothing is being blacklisted by mistake - in which case you should remove it from the blacklist. (You can read the messages in the message centre so do you really need them delivered?) It sounds to me that you should be using a filter in your mail client to move the messages to a separate folder and not miss-using the blacklist feature.

I think you have to remember something. Plusnet probably have around 300,000 users and are using what Postini consider legacy features. To quote from the Postini website:


If a relatively small company like Plusnet goes to a giant like Postini it is largely a case of "take it or leave it".

IIRC, this is the first time I've had a message spoof a plusnet sending address that wasn't my own. (Other people may be worse off than me, spam-wise!)  Pleased to see it was handled correctly by plusnet.

What on earth is going on with postini at the mo ?  Messages are frequently (if not always) taking SEVERAL MINUTES to be received by the postini servers. Is this "by design" or "by c*ck-up" or "by act of god" or what ?

The following is a netmeter graph simply because I happened to have it handy whereas I don't have Mr.Laptop with Ethereal on it handy at the mo but it clearly demonstrates the problem anyway:



A = 2 line text message sent via a non-postini SMTP server to my A/C
B = The initiation of the same message sent via postini to my A/C
C = The completion of the same message sent via postini to my A/C
D = The same message sent via a non-postini SMTP server to my A/C

The 'blips' are regular messages from one of my routers and would be at approximately 10 second intervals. This, of course, also stacks up in this instance with the report from my mail program showing the time taken to send the message.

The length of time postini is apparently holding the connection open but stalled before completing the transfer is several minutes. 2 mins is very typical but I've seen around 8 mins on some attempts and a maximum of around 15 mins so far. This simply cannot be right and will no doubt be causing no end of problems for anyone attempting to send any significant number of messages to various PN customers.  I don't know exactly how long this has been going on for but something tells me perhaps around 2 weeks because that's when I first started seeing other very strange things happening on all my monitored postini A/Cs 

Who the h*ll has fiddled with what because this is a relatively new issue.  Seconds are fine and perhaps to be expected but minutes are just crazy especially when you get well into double digits thereof !!!

Messages to me from external or Plusnet SMTP servers and received via Postini and Plusnet POP3 servers have taken 3 or 4 seconds in the last few days.

So nothing appears to be going on from my point of view.

Not quite the same because I'm not talking about the time taken for ME to actually receive the message at MY A/C although there clearly is (or rather should be) a direct relationship with that time of course.

What I'm saying is that postini accepts the connection request almost immediately but then appears to ignore it and does not respond again for several minutes. This means that from establishing a connection to postini to actually completing the transfer of a very small message takes several minutes. It does not mean that postini didn't respond at all for several minutes because that would have generated a timeout at my end pretty quick, probably in a few 10's of seconds. The end result is a connection tied up for several minutes for absolutely no good or sensible reason.  It very much sounds to me like postini have intentionally changed something in an attempt to deter, confuse or generally screw up spamming activity or at the very least slow them down considerably. In principle this could be just fine of course if done sensibly and reliably (and I've suggested implementing similar tactics many times before) but in practice I can imagine LOADS of problems at the moment for totally legitimate senders who just so happen to send a lot of totally legitimate messages over a short space of time to PN customers. 

Despite my lack of anything much in the way of real evidence, I think my money is very much on a postini change implemented around a couple of weeks ago because that stacks up with other very strange things happening around that time and perhaps it also stacks up with the well known about problems with mail being lost from certain other sources    I also note that despite requests from at least one customer for a full and detailed explanation of that problem explaining the reasons for and resolution of said problem, nothing has yet been forthcoming from PN.

15 minutes to complete the sending of a totally legitimate 2 line text message ? Someone's taking the p... having a giraffe 

I also note that postini (or more likely PN) aren't exactly being honest with the timing quoted in message headers either. However, I suspect this could well be the usual problem with the various PN server clocks not being synchronised to a time server so times can (and often do) differ considerably depending on which server you happen to get connected to.

Typical headers from one message sent around the time I posted my previous comment but only just received now:


Times quoted in headers are *very* dubious to say the least because this particular message most certainly wasn't available for DL via pop3 sometime around 0200 hours !!  I have also just received several others sent around the same time that should have been available well before 0200 hours.

Why is postini establishing a connection then stalling for a completely unreasonable period of time on some/most/all occasions ?

Maybe http://en.wikipedia.org/wiki/Tarpitting ?

B.

Yup, could well be, because that's exactly the sort of thing I'm on about and pretty much what I've suggested many times previously without actually knowing there was some history of already doing it.

The key comments in the (not necessarily entirely correct) Wiki page intro are delays in the order of SECONDS, variable delays subject to message LENGTH/CONTENT/SPAM_RATING, such systems CAUSING PROBLEMS with regards to mail lists etc. and the implied need for EXCESSIVE WHITELISTING.  It has to be said that there is absolutely no evidence that whitelisting helped in any way here of course for the example message I quoted.

This very much appears a new and as far as I can tell totally undocumented 'feature' of postini but it most certainly goes a very long way towards explaining the major problems that I have always experienced with loss of totally legitimate regular mail from some sources.

The example message I quoted previously is clearly not spam according to any of the available postini indicators and was very short so any delays introduced by any means should have been pretty negligible. There is absolutely no documented or even loosely plausible reason why it should have been treated in this way at all let alone why it was apparently intentionally dragged out for around 15 minutes.  If this is now (and perhaps has always been) standard postini practice then it is in no way surprising to me that there are so many problems with totally legitimate genuine mail going mysteriously AWOL.

So come on PN: Where is this 'feature' documented ? Why was this particular message 'caught' ? How long has this been going on ? ... or is it perhaps all speculation and just one of those things that sometimes happens because someone screwed up somewhere ? Definitive answers would be nice, preferably here and real soon rather than on the proverbial postcard subject to Royal Mail loss or delays etc. 

It seems that Bob and co now prefer to run for cover instead of answering any queries raised in this thread.