8
Build 8025 Feedback
Question asked by JerseyConnect Team - 12/22/2021 at 6:51 AM
Unanswered
As the title suggests I'm just starting a thread to collect feedback on the latest release. 

There is at least one bug fix we need, nothing urgent, but due to the holidays we won't get this into production for a few weeks. I do plan on trying to get this tested in a lab, so I'll report any issues. 

Anyone else have an urgent need to get this installed or had a chance to try it out?

64 Replies

Reply to Thread
20
Tim Uzzanti Replied
Employee Post

Build 8025 may be one of the best SmarterMail Builds we have ever released. We spent that last couple months of 2021 evaluating the environments of about 25 companies, distributing custom Build after custom Build to them, and improving their overall performance and stability until we literally saw nothing else we could improve.

The reason we did this is because SmarterMail saw significant user growth in 2021, primarily from Enterprise organizations trying to leave Microsoft Exchange due to various vulnerabilities and the direction Microsoft is heading with Microsoft 365. Many of these companies said they didn't want to spend large sums of money upgrading older Exchange environments (software and hardware) and didn't want the increased costs of Microsoft 365, etc.

Prior to these large Enterprise moving to SmarterMail, many of our MAPI users were smaller companies that have the ability to easily make these moves and/or hosting companies, service providers or ISP's enabling MAPI on their SmarterMail servers for existing customers who want to pay a premium for MAPI/EWS, etc. In most cases, these servers see gradual growth with new features or functionality in SmarterMail.

But, with the recent additions of enterprise organizations enabling 1,000 or 10,000 MAPI/EWS or EAS accounts overnight, we started seeing issues that we could only see once in a while with other customers.

These issues weren't being caused by normal activity, but we discovered SmarterMail did not gracefully handle clients that were in bad states. For example, we do not support Microsoft Outlook 2013 because Microsoft does not update it anymore and it has a number of severe issues that can cause Outlook 2013 clients to go bad and almost DoS a server.

Well, when you have 1,000+ clients (Outlook 2013 or higher) that have sync keys or data from previous Exchange versions and these clients/accounts don't get deleted and re-added to connect to SmarterMail as a new connection, Outlook will use older sync keys and information related to the previous Exchange against SmarterMail. That breaks the protocol. Although we did anticipate this happening and/or clients breaking for unkown reasons, we had no idea how many ways it could attack SmarterMail and cause issues for other users.

The improvements found in Build 8025 are the direct result of companies absolutely destroying SmarterMail servers with incredible amounts of bad data and bad clients. We also implemented logging for companies to look at MAPI and EWS clients that may be improperly configured and are hitting the servers in ways that are now handled MUCH better, but may be stuck because the Outlook client was working with a different server and needs to be re-connected and re-added with the new server so that they have all the proper sync information.

There are a number of other fixes and improvements in this build as well.  We spared no expense or time to find absolutely everything we could improve.  SmarterTools technical support is already at an all time low due to how stable our products have been and results we made at the end of this year, should give us even more bandwidth.

Happy Holidays Everyone!
Tim Uzzanti CEO SmarterTools Inc. (877) 357-6278 www.smartertools.com
1
Scarab Replied
Although we've been using SmarterMail ever since the divergence/exodus from iMail (it was so long ago...was that 2004ish?) and used to be early adopters but with v17/100X we've been finding ourselves holding off on updating SmarterMail for a couple of weeks after a new update comes out to see where the chips fall (since our servers were IOPS limited and we were limping along as it was, especially with exceptionally long Write Queue Wait Times, that we couldn't risk any changes making it worse until we migrated to new physical servers with significantly higher IOPS).

This update fixes 7 specific issues that we've been struggling with lately, even after migrating to better bare-metal servers with multiple SSD RAID arrays, so we are going to take the plunge and update immediately out-of-band after peak-hours are over. This update really does feel like a holiday present from SmarterTools.
1
Tim Uzzanti Replied
Employee Post
The developers will glad to hear that.  Ran them pretty hard before the holidays. 

Healthy and Happy Holidays to all!
Tim Uzzanti CEO SmarterTools Inc. (877) 357-6278 www.smartertools.com
1
echoDreamz Replied
Amazing... Great work Tim and team! Hopefully... Maybe this new build will help with complaints we get from O365 and Exchange migrated customers about ActiveSync folder contents being out of sync across devices. This has been an issue we've had complained about for years.
0
Andrea Free Replied
Employee Post
We're really hoping so as well, echoDreamz! If protocol problems persist on Build 8025, we'd ask that you remove the account from the device, deleted the Synchronized Devices entry, then re-add. This will allow us to have a fresh look at the connection and guarantee any old misconfigurations are no longer at play. If the issue persists, we're happy to work with you in a support ticket to get to the bottom of this. 

Happy holidays! 
Andrea Free SmarterTools Inc. 877-357-6278 www.smartertools.com
2
Sébastien Riccio Replied
We're currently running the custom build (8020) and no issues were reported since. I guess it should be safe to go 8025 and will be updating to it soon.
Happy holidays to everyone.
Sébastien Riccio System & Network Admin https://swisscenter.com
2
echoDreamz Replied
Same, we've been on 8020 for maybe a week now, no issues to report.
1
John Marx Replied
Everything is good here.We had a couple where MAPI weren't working and now are so very happy to close those tickets on our CRM as well.
2
Tim Uzzanti Replied
Employee Post

Shaun,


I can't find any active tickets for you, but I did review the few tickets you've opened over the last 3+ years. This includes the one a couple months ago relating to the corruption. I remembered this ticket because it was unique and I personally wanted to take a look at your server. It's one thing to have corruption on the main json but odd that the tmp file process and recovery would also be corrupted.

In any case, logs do provide various exceptions for corruptions or accounts that can't load, etc. In your case, we even provided you a script and a list at the time of 99 accounts affected so you didn't need to do it.

If you're still having issues with your server, please start a ticket. I would want to know what you mean by a "stall" because that would be very different that the things we were seeing. What we did see from time to time was a stack overflow issue which was a nightmare and it was like one email out of millions that could cause it. We also had resources ramping up mostly in high volume environments and primarily with email clients that are misconfigured. This could cause unusual behaviors after awhile because the .NET Framework and garbage collection could pause our process for periods of time while it tried cleaning up handles and memory.


Also, if you want to be proactive with MAPI and EWS clients that could be misconfigured or not running well, please check the admin logs.
Tim Uzzanti CEO SmarterTools Inc. (877) 357-6278 www.smartertools.com
0
Stefan Mössner Replied
Hi all,


Kind Regards and merry christmas
1
Heimir Eidskrem Replied
Installed last night and so far no issues.
Its not been long and its a slow day but so far so good.
It did fix the DKIM issue we saw with yahoo it seems.
Or I should say the very little testing I have done so far shows that DKIM is now recognized by yahoo.

5
Tim Uzzanti Replied
Employee Post
Stefan,

The comment scared me because we haven't had SMTP issues in a long while beyond DKIM which is kind of a moving target. When I talked to the guys this morning, they remembered your issue because they went out of our way to look at the custom code and demonstrated it wasn't SmarterMail.

So that everyone is aware, the component is prematurely closing the connection while we do SPAM checks at the SMTP level and when we respond back to the SMTP session that all is good to continue, the component had already disconnected.  The log Stefan pasted is an example of this, when a client disconnects before we respond back with all good.  

Kind of surprised by your post after after all our effort and communication with you.

The fix for you is the component you are using.
Tim Uzzanti CEO SmarterTools Inc. (877) 357-6278 www.smartertools.com
3
echoDreamz Replied
Stefan,

We had an issue similar to this with an app we built in-house, turned out to be the Mailbee SMTP component that was the issue, they fixed it up and we worked for years before moving to the open source MailKit project.
2
Stefan Mössner Replied
Tim,

I have to disagree. Neither was there a communication with me, my linked post was never answered because there were other posts with different issues and you only answered these posts. Nor I can't see an issue with the application trying to send a message because the issue starts with SmarterMails build 7978. And I'm using the same application on another server with sending the message to a Microsoft Exchange server without any issues.

Here's the complete SMTP log of the try to send the message:

00:09:40.731 [aaa.bbb.ccc.ddd][42060648] rsp: 220 xxx.net
00:09:40.732 [aaa.bbb.ccc.ddd][42060648] connected at 27.12.2021 00:09:40
00:09:40.733 [aaa.bbb.ccc.ddd][42060648] Country code: de
00:09:40.735 [aaa.bbb.ccc.ddd][42060648] cmd: EHLO XXX
00:09:40.735 [aaa.bbb.ccc.ddd][42060648] rsp: 250-xxx.net Hello [aaa.bbb.ccc.ddd]250-SIZE 41943040250-AUTH LOGIN CRAM-MD5250-8BITMIME250-DSN250 OK
00:09:40.736 [aaa.bbb.ccc.ddd][42060648] cmd: AUTH CRAM-MD5
00:09:40.736 [aaa.bbb.ccc.ddd][42060648] rsp: 334 XXX
00:09:40.736 [aaa.bbb.ccc.ddd][42060648] Authenticating as user@xxx.net
00:09:40.737 [aaa.bbb.ccc.ddd][42060648] rsp: 235 Authentication successful
00:09:40.737 [aaa.bbb.ccc.ddd][42060648] Authenticated as user@xxx.net
00:09:40.741 [aaa.bbb.ccc.ddd][42060648] cmd: MAIL FROM:<sql@xxx.net> SIZE=25994
00:09:40.741 [aaa.bbb.ccc.ddd][42060648] senderEmail(1): sql@xxx.net parsed using: <sql@xxx.net>
00:09:40.741 [aaa.bbb.ccc.ddd][42060648] rsp: 250 OK <sql@xxx.net> Sender ok
00:09:40.741 [aaa.bbb.ccc.ddd][42060648] Sender accepted. Weight: 0. Block threshold: 30.
00:09:40.741 [aaa.bbb.ccc.ddd][42060648] cmd: RCPT TO:<user@xxx.net> NOTIFY=FAILURE
00:09:40.744 [aaa.bbb.ccc.ddd][42060648] rsp: 250 OK <user@xxx.net> Recipient ok
00:09:40.759 [aaa.bbb.ccc.ddd][42060648] cmd: DATA
00:09:40.759 [aaa.bbb.ccc.ddd][42060648] Performing PTR host name lookup for aaa.bbb.ccc.ddd
00:09:40.792 [aaa.bbb.ccc.ddd][42060648] PTR host name for aaa.bbb.ccc.ddd resolved as server.yyy.de
00:09:40.794 [aaa.bbb.ccc.ddd][42060648] rsp: 354 Start mail input; end with <CRLF>.<CRLF>
00:09:40.794 [aaa.bbb.ccc.ddd][42060648] senderEmail(2): sql@xxx.net parsed using: "SQL Backup Server" <sql@xxx.net>
00:09:40.795 [aaa.bbb.ccc.ddd][42060648] rsp: 250 OK
00:09:40.796 [aaa.bbb.ccc.ddd][42060648] Client socket is disconnected! Disconnect exception encountered: False, IsDisconnected: True
00:09:40.796 [aaa.bbb.ccc.ddd][42060648] Received message size: 25997 bytes
00:09:40.796 [aaa.bbb.ccc.ddd][42060648] Successfully wrote to the HDR file. (D:\Mailserver\Spool\SubSpool4\656948297749.hdr)
00:09:40.796 [aaa.bbb.ccc.ddd][42060648] Data transfer succeeded, writing mail to 656948297749.eml (MessageID: <3a2b0b379f6e2dbe992c8584525e91b8@server>)
00:09:40.796 [aaa.bbb.ccc.ddd][42060648] disconnected at 27.12.2021 00:09:40

Looking at the delivery log I can see that the message isn't delivered to the SpamCheck queue:

00:09:42.943 [48297749] Delivery started for sql@xxx.net (via user@xxx.net) at 00:09:42
00:09:42.943 [48297749] Removing Spool message: Killed: False, Failed: True, Finished: False
00:09:42.943 [48297749] Delivery failed for sql@xxx.net at 00:09:42    [id:656948297749]

The spamChecks log doesn't show any lines regarding this message.

So the issue happens before the spam checks by SmarterMail.

Kind Regards
10
Scarab Replied
So far we've been very pleased with Build 8025. No bumps, just lots of pleasant fixes.

@Tim Uzzanti

We infrequently but regularly (once a week to once a month) experienced the same problem Shaun Peet is reporting with individual users and entire domains having their settings.json corrupted due to a brief crash/restart of SmarterMail (generic ".NET clr.dll encountered an error") for the past year+, although we haven't had the issue since installing Build 8025 (knock on wood). It is nice that SmarterMail now logs the corruption, but we've had it a bit easier identifying these as the Waiting to Deliver Queue on our Incoming Gateways will start piling up with emails for the corrupted users/domains so we setup a Zabbix trigger to alert us when that count gets unusually high so that we can restore the latest archive and reload those domains hopefully before a customer contacts us.

When a domain or user gets corrupted the domains will show a Disk Usage of 0.0/0.0 MB (0%) and users show a X.XGB/100MB (so it shows their actual disk usage but always with a limit of 100MB). The server won't allow the users to authenticate via IMAP/POP/SMTP/Webmail and any attempts at email delivery to the corrupted domains/accounts results in "rsp: 550 No Such User Here". The domain or user card in the SmarterMail Administration panel will give an error about being unable to load resource.

We never reported it previously because we just assumed it was due to too high of a Current Disk Queue Length resulting in data corruption when SmarterMail shutdown inelegantly (as the users with corrupted settings.json seemed to be the ones with mailboxes being indexed at the time it crashed), so we focused on upgrading our hardware to accommodate higher Disk I/O and IOPS (went from a Hyper-V instance with Raid-5 HDDs running Win2K19 to bare-metal with two Raid-10 arrays with SSDs running Win2K22).

Ultimately it would be nice if there was some kind of built-in check at SmarterMail startup if a crash was detected to look for corrupted users and domains and automatically restore those from last archive before loading. If there was an inelegant shutdown of SmarterMail I'd be okay with a longer load time on the next startup rather than dealing with corrupt users and domains.
3
Shaun Peet Replied
If I could give a million upvotes to those last two posts, I would.
0
Vahn Babigian Replied
After 8025 update accounts setup with IMAP in older Thunderbird versions stop working. Thunderbird version 52.9.1 and older. (Cannot update Thunderbird as these are legacy PCs running Windows XP). 
3
echoDreamz Replied
Ouch... time to let those go :)
3
Heimir Eidskrem Replied
I woudlnt even mentioned that something didnt run on something as old as XP.
No excuse for that.  No way XP should have access to the internet.

0
Paul Blank Replied
Agree totally on the XP front. Windows 10 is surprisingly stable these days, and the 32-bit version works very well for most 16-bit legacy apps - in a DOS box or Windows app. And for whatever reason, Win10-x86 (32-bit) feels a bit zippier than x64, as long as you can work within the 4GB RAM limitation of x86. And refurb'd/recycled PCs - you can get them so inexpensively -  fitted with boot SSDs, even several years old*, typically run Win10 very well, x86 or x64. YMMV, of course.

*Example: HP Elite 8000 with Core2 Duo CPU and 4GB RAM, first released ca. 2009 (!)
1
Kailash Aghera Replied
Any solution for the following error in SMTP logs:

Client socket is disconnected! Disconnect exception encountered: False, IsDisconnected: True

I have also noticed this error in SMTP logs. It is not appearing for all but it is affecting many users. What is the cause of the above error in SMTP logs? I have using latest build Build 8025.

Thanks!
3
FrankyBoy Replied
URGENT - URGENT - URGENT

Hello, seems to have a bug with Outlook 2019 - MAPI and Calendar appointments. If we create an appontment from the webmail, Outlook Mac (EWS) or iPhone (Activesync), This appointment will not appears correctly in Outlook 2019.

All still ok on the Mac, the iPhone and the webmail, but Outlook 2019 do not take care of the UTC -5, which is our Time Zone - Will IMMEDIATELY open a ticket for this.
2
Michael Replied
Yup. Same issue. Every user effected. We're in PST and every appointment sent by or received from Outlook user is showing +8 hours. We have an open red-hot support ticket since the weekend. I believe ST is working on a fix now. As folks come back from holiday everyone is noticing. Many are missing meetings. 

If you haven't updated to 8025 yet and you rely on meetings and Outlook MAPI integration, I'd hold off on updating. This time zone issue is causing chaos for us.
3
Proto Replied
I really appreciate the Michael and FrankBoy posting here in addition to opening the tickets.
I was getting worried a couple of days ago but this has us holding off on the update for sure.

These are the kinds of problems that have cost us several thousand users over the years.
Companies have patience for most things but when meetings get missed, resource schedules overlap or we end up with changes to the base of email that take many hours to clean up they loose patience.

Not bashing ST here, We wouldn't keep supporting it if we didn't like where it's going.
It would be nice if, in addition to the exhaustive, protocol, new feature, file system and other testing that takes place, there was a short "high perceived value"  list that was retested before each and every release to reduce the number of times the inevitable slip though.

For corporate accounts the Outlook functionality is still critical and the parts of it that ST provide need to be flawless to avoid a backlash.

Thanks again for the heads up!
SmarterMail(tm) MAPI over HTTP - Let's flesh it out for Outlook with a full set of Exchange like features!
5
Tim Uzzanti Replied
Employee Post
Already have a fix for the calendar issue and is in QC.

Regarding the SMTP issue, we are going to remove the need to verify the client sees are OK and will send the message.  The clients are the problem (most often components used for mass mails) but we are not going to fight the fight and as a result, there are circumstances a duplicate message can occur. This is also in QC.


Tim Uzzanti CEO SmarterTools Inc. (877) 357-6278 www.smartertools.com
1
Chris Daley Replied
So 8025 changed how SM handled SMTP clients that didn't issue the quit command?
0
Matt Petty Replied
Employee Post
It was specifically checking if the client was still connected when we issued the "250 OK" for the DATA command. So if the DATA command took a little while (spam checks) and the other server disconnected before* we send "250 OK" if we detect they are disconnected we right now with 8025, will NOT accept the message and instead treat it as a failure.

Pre-8025: We would issue "250 OK" and not care if the client was still there and we would accept the message into the spool.

We will be adding a new setting that will allow you to disable this new behavior.
Matt Petty Software Developer SmarterTools Inc. (877) 357-6278 www.smartertools.com
2
Chris Daley Replied
Have you tested how a .net framework v2/3.5 client will handle the original 8025 implementation? The smtp client in .net 2/3.5 does not issue the QUIT command, there is no Dispose method as you have in .NET 4+ so the way the clients work around it is by setting SmtpClient.ServicePoint.MaxIdleTime to a low value which would create the scenario SM can no longer handle. Setting the max idle time forces .net to release the socket.

Essentially the 8025 implementation was a breaking change, note... you haven't even mentioned this 'improvement' if you can call it that in the release notes?

The logging for this in SM is also poor as support are aware, the email literally vanishes with no indication in the log as to why, it should have said 'connection closed prematurely'

You should have listed this in the release notes and provided it as a setting in SM itself that way admins could switch back if need be. You have no way of knowing how many customers / applications are effected by this due to how SM handles it in the logs.
0
Matt Petty Replied
Employee Post
This behavior doesn't deal with the quit command. It specifically is about when the client disconnects in regard to our "250 OK".

Also when this behavior occurs we print
"Client socket is disconnected! Disconnect exception encountered: false, IsDisconnected: true"
into the SMTP log.

After we add this setting, it will now print
"Client socket is disconnected! Disconnect exception encountered: false, IsDisconnected: true, This message will be rejected."
OR
"Client socket is disconnected! Disconnect exception encountered: false, IsDisconnected: true, This message will still be accepted."
Based on the setting.
Matt Petty Software Developer SmarterTools Inc. (877) 357-6278 www.smartertools.com
2
Tim Uzzanti Replied
Employee Post
The scope of this Client Disconnect / SMTP issue is important.  

We have had less than 10 tickets.

How SmarterMail works today will continue to be the default because it is correct.

There will be a setting at the SMTP IN level allowing clients that disconnect to continue to send a message.  

If you want to call this a win, call it a win. You will be opening yourself up to duplicate messages and more. 

The clients that are most often an issue are components or tools used for mass emails.  
Tim Uzzanti CEO SmarterTools Inc. (877) 357-6278 www.smartertools.com
0
Sabatino Replied
sorry i wanted to wait to install the 8025. i was almost convinced to install when i saw the post about thunderbid on xb. although we are trying to convince users to abandon xp we still have several users with xp and tb. do you confirm the problem with imap? will it be fixed?
Sabatino Traini Chief Information Officer Genial s.r.l. Martinsicuro - Italy
6
Chris Daley Replied
Tim,

Breaking changes should be communicated and you should have made it a setting in the first place.

.NET 3.5 is supported by MS until 2028 so there will be legacy apps out there that will be implementing the same workaround for the smtpClient bug I provided info on. This isn't just about 'bulk email'.

If SmarterMail is for the enterprise as you claim you will have to deal with legacy apps.

10 tickets may not sound like a lot but considering the holiday break not everyone will have updated and not everyone will have noticed the issue yet. You have no way to tell how big of an impact your decision has made.

Looking back at your first post on this thread "Build 8025 may be one of the best SmarterMail Builds we have ever released." I would disagree with that statement, it may be good for MAPI but not for the core.

If every developer in the world followed standards (RFC's etc) the world would be perfect (from a development perspective) unfortunately they don't, MS especially. In the future you need to consider that before making decisions.

I'm not going to argue with you publicly any further on this but I believe others will support my view.
6
Tim Uzzanti Replied
Employee Post
Chris, are you really trying to lecture us about RFC's and balancing clients and functionality?  If so, you really have no idea what we do all day. 95% of what we do is work around broken clients including Microsoft.  When a statistical anomaly of your customers have an issue with an occasional third party mass mail component that should be replaced because we are telling you its broken, I'm not really interested in having a conversation with that person anymore.  Just like, I'm not going to have a conversation about Windows XP anymore.  Ironically, the company actually implemented a setting for that statical anomaly.

I could really care less who is a customer or not.  What I do appreciate, are customers who appreciate the work we put in.  We don't build product to make money, we build product to help people.  

Kind of sick of fighting people about shit they should spend a little time upgrading for fixing.  We deal with people not updating 15 year old shit or think no matter what client, hardware or scenario SmarterMail should just work.  Sometimes common sense just escapes people.

Call me frustrated because, we really do give a crap and we really do try to help people and companies run as well as possible and why the CEO (myself) has been one of the most active in the communities through the holidays.  But, we can only help the people who want to help themselves.
Tim Uzzanti CEO SmarterTools Inc. (877) 357-6278 www.smartertools.com
0
Chris Daley Replied
I'm not trying to lecture you; I'm simply asking you to say something when you make breaking changes or make it optional. My hands are tied, what do I say to customers still running .net 3.5 which is a supported product?

Its not a personal attack it never has been. 

You really need to take a break.

2
Paul Blank Replied
"Aren't standards wonderful? There are so many to choose from." 

Alas, Microsoft is the 800 lb. gorilla in the room, and they remind us of that every day... see Windows 11, for example.
4
echoDreamz Replied
@Sabatino that would be their problem IMO... No excuse to still be running XP in 2022. The sooner we stop recognizing and supporting these clients, the better. 
2
Kyle Kerst Replied
Employee Post
I agree 100% @echodreamz. When I was running my own shop I let XP/Windows 7 clients and temporary passwords slide long enough that one of my customer's was nearly compromised because of it. At that point I put my foot down - if I'm going to be responsible for it - it needs to be running versions that don't have known security and stability issues. 
Kyle Kerst System/Network Administrator SmarterTools Inc. (877) 357-6278 www.smartertools.com
0
Sabatino Replied
I would like a world without xp but unfortunately now it is not like that.
We need to analyze the context first. I have a mail server and I sell e-mail services. Often I don't have any other customer relationships than this, so I'm not the one who takes care of the hardware and systems. So blocking some of them overnight for me means losing the customer. The real problem is not knowing who is using tb on xp first. If there were in the imap log, I would be really happy, I would first intervene on the customer.

Also no one answered me. By installing the 8025 all the tb on xp in imap will have problems or only in special cases?
Sabatino Traini Chief Information Officer Genial s.r.l. Martinsicuro - Italy
2
echoDreamz Replied
Sabatino, that may be true, but the sooner companies start saying "Oh, you run XP? No support. Sorry." the sooner we can push people off of XP. Keeping this "Oh, XP exists, so we just deal with it" only extends the problem.

XP went EOL nearly 8 years ago, no excuse to still be running it, especially with internet access. And it's not overnight, XP has been dead for a long long time, users on it, know it and are just too stubborn, lazy or cheap to move on from it.
3
Zach Sylvester Replied
Employee Post
Hello Everyone, 

Just following up. We have a custom build that should fix the MAPI calendar issue as well as the SMTP disconnect issue. To get access to this custom build please open a ticket with us and we would be happy to provide it.  Have a wonderful weekend everyone. 

Kind Regards, 
Zach Sylvester System/Network Administrator SmarterTools Inc. (877) 357-6278 www.smartertools.com
0
Paul Blank Replied
I suppose there are legacy apps that "only" run under XP. And some companies are too cheap or too poor to try to get their app running anywhere else. 

The worst I have to deal with these days is a couple of clients with old FoxPro LAN (DOS!) apps that thankfully run well in DOS boxes on Windows 10 x86 or reasonably well in vDOSPlus on Win 10 x86 or x64.
0
Kailash Aghera Replied
I can confirm that nopCommerce 4.10 and 4.20 are affected by SMTP change. I believe it should be related to SmtpClient but there is not way we can fix this and we cannot force the client to upgrade it immediately to latest version (in which they are not using MailKit). Many themes, plugins etc. may not be compatible and it is not possible to upgrade overnight.

This is just an example script and there may be large number of scripts/code affected  by latest build. Instead of requesting custom build. It is should be available in general availability or at least allow an option to reverse this.
0
Sabatino Replied
Apart from the discussion xp yes xp no I still do not have a certain answer

Will all xp's with thundebird in imap stop working with the upgrade to 8025?
Sabatino Traini Chief Information Officer Genial s.r.l. Martinsicuro - Italy
3
Sébastien Riccio Replied
After reading about the worrying MAPI date issueI decided to do some tests on some customers accounts we were pushed by SM to migrate from EAS to MAPI.

Damn, they are also afected:

Here is one entry in the webmail:


Here is the same entry in Outlook (I'm testing with Outlook 2021 LTSC):


There is 1 hour difference (Outlook shows it one hour before the webmail).

We had no complaints yet as most of these MAPI users are still on holiday. But when they will be back monday, that will be catastrophic. If they are missing appointments we are DEAD! How can we trust the product for professional use ...

Where can be downloaded the build for fixing this ? Why is it not publicly available and need to open a ticket as this is CRITICAL.

edit: opened an own thread for the issue. It's too important to be lost in a mixed stuff long thread:
https://portal.smartertools.com/community/a94563/build-8025-critical-mapi-calendar-date-issue-again.aspx

Sébastien Riccio System & Network Admin https://swisscenter.com
0
Stefan Mössner Replied
Zach,

I have the SMTP related issue with the newest version of SQL Backup Master. This software is still under development, so there are frequently new versions. It's not an old software tool that can't be updated anymore.

I'm a user of the free edition of SmarterMail (home use). So I cant open a ticket. How can I get the custom build to test the SMTP setting fixing the issue?

Kind Regards
0
David Fisher Replied
Hi,

  Since SmarterTools is off for the weekend, anyone that needs the 1/6/2022 custom build PM me.

-dave
1
Zach Sylvester Replied
Employee Post
Hello, David

 I have sent you the custom build. 
I will be checking tickets periodically through the weekend for custom build requests. 

Regards,
Zach Sylvester System/Network Administrator SmarterTools Inc. (877) 357-6278 www.smartertools.com
0
Zach Sylvester Replied
Employee Post
Hey Stefan, I messaged you. 


Zach Sylvester System/Network Administrator SmarterTools Inc. (877) 357-6278 www.smartertools.com
0
Sébastien Riccio Replied
Hello David, thank you very much for the link. 

Kind regards
Sébastien Riccio System & Network Admin https://swisscenter.com
3
Sébastien Riccio Replied
David, may I ask if you were given a changelog with the build. Is it safe to downgrade in case of things get worse ? Just trying to be sure before installing as some builds sometime aren't downgrade friendly.

Thanks a lot.
Sébastien Riccio System & Network Admin https://swisscenter.com
0
David Fisher Replied
Sebastien,

  I believe it is only the two changes, no changelog provided, you should be able to downgrade, the last non downgradable one was build 7879 dated July 28, 2021 when they went from stat3 to stat4 on the statistics files.

-dave
0
Zach Sylvester Replied
Employee Post
Hello Sébastien,

While we don't recommend downgrades. There are only two changes made with this build and it shouldn't be changing the files in a way that makes it impossible to downgrade. 

Kind Regards, 
Zach Sylvester System/Network Administrator SmarterTools Inc. (877) 357-6278 www.smartertools.com
3
Sébastien Riccio Replied
Zach, it's a bit hard to hear that downgrade is discouraged when installing a new build is often like playing russian roulette... 
What are the other option in case things goes wrong ? a complete restore of a pre-update backup (10TB+ in our case ?).
Even with snapshots on our storage system which can be easily rolledback, we would lose our customer new content between the pre-update snapshot date and the rollback...

That being said. I've installed the provided custom build.

While it seems it fixed the timeshifting issue between outlook(mapi) and webmail for NEW appointments, previously added appointments still shifts... 

After the update I took a look again at a customer calendar and it looks like a huge mess. Some entries are correct in both webmail and outlook, some are shifted in outlook, some appointments overlap other appointments in outlook (but in webmail they arent overlapping).

I now fear we are completly losing original and correct information about the appointments date/time. How to know what date can we trust. Webmail, Outlook ? How to fix the issue for the post-update added appointments...

On monday if our customers are missing appointments because of these issue we're going 
to face huge troubles.


Sébastien Riccio System & Network Admin https://swisscenter.com
0
Stefan Mössner Replied
Hello Andrea, David and Zach,

happy new year to all of you.

Thank you all for sending me the download link for the custom build 8041. I installed it. It took a long while to install it (longer than the builds before).

After enabling the setting "Continue delivery if session is disconnected by client" in Settings > Protocols > SMTP In card I started a manual SQL backup with SQL Backup Master to see if the message will be delivered. I can confirm that the status message of SQL Backup Master is now delivered. Now I can also see that the message is passing the anti virus and spam checks. I think this will work now for the scheduled backups, too. Thank you for implementing this feature in SmarterMail.

Kind Regards
1
Thomas Lange Replied
I installed custom build 8041 on our Windows Server 2012R2 and I noticed an unusual very long installation/setup process with this custom build (like Stefan already mentioned). The screen/message with "Copying new files..." showed up for around 15 minutes.... and task-manager shows CPU aroung 50% during that time...., prior to setup the custom build I rebooted our server, deinstalled/removed buidl 8025, rebooted the server again and then I started the setup.exe of custom build 8041.
0
Michael Replied
Custom build 8041 resolved the appointments issue described above. But agree, the install was unusually long. 10 or 15 minutes in the Copying files step. Wonder why?
0
FrankyBoy Replied
Same thing for us in terms of installation time. It's not a big deal except that I thought for a while that the installation had crashed, but with a little patience, everything went well.
3
Sandro Replied
Hi all

I have an important question before updating:
You, Sébastien Riccio, wrote the following about the time shift issue after getting the custom build:

While it seems it fixed the timeshifting issue between outlook(mapi) and webmail for NEW appointments, previously added appointments still shifts... 

After the update I took a look again at a customer calendar and it looks like a huge mess. Some entries are correct in both webmail and outlook, some are shifted in outlook, some appointments overlap other appointments in outlook (but in webmail they arent overlapping).
Does someone know if this (mix with wrong old and correct new appointments) is only a problem when updating to 8025 and than updating from 8025 to custom build 8041? 

Currently we are running 7887 which worked very well for us but since we will migrate some new customers to SmarterMail in the next few days, I wanted to update before migrating them to make sure they set up everything on their end using the most current version of SmarterMail.

3
Stefan Mössner Replied
Hi all,

the installation process of the now public available build 8048 doesn't take so much time any more as the previous custom build 8041 took.

Kind Regards 
1
Webio Replied
I just wanted to let you know that I've upgraded from 7957 to 8055 and my clients started to experience this issue with "This message will be rejected" error. At friday it was 1010 occurences in SMTP log and today it was about 350.

It is hard to force users to upgrade their apps in shared hosting scenario business. I've suggested "MaxIdleTime" param but still no one knew about this could happen. After client ticket I've started to dig in and found this topic for 8025 build.
3
Sébastien Riccio Replied
Hello Webio,

If i'm correct this seems to be related to the change at how SmarterMail handles the incoming mail when the remote server disconnects right after submitting the mail DATA (and does not wait for confirmation from the server).

AFAIK this is an unusual behavior mainly used by spammers or mass mailers that want to spare time.
This especially to optimize deliveries to servers that are doing mail content scanning right after the mail DATA has been submitted and before returning the result.
This also mean they don't care if the message is accepted or rejected.

After some disucussion in this thread SmarterMail added a setting to switch back to the previous behavior (allow a mail sent by a remote server that disconnected early) like Matt explained here.

The setting is disabled by default. If you feel it's an issue for you, you can enable it in the SMTP-in card of the protocol settings to go back to the previous behavior.


We left it disabled on our system. I checked our logs and we have only a very few of these from clients submitting mails for delivery to port 587.
I guess this is because all incoming mails are routed through our filtering appliances first, which then relays the mail to the SM server without any eary disconnect.

I hope this helps.
Sébastien Riccio System & Network Admin https://swisscenter.com
1
Sébastien Riccio Replied
Hello Sandro,

Does someone know if this (mix with wrong old and correct new appointments) is only a problem when updating to 8025 and than updating from 8025 to custom build 8041?  
As far as I understand, the time shifting issue with MAPI was an issue introduced somewhere around build 8025 and was then resolved.
I guess a straight update from your current build to the latest would not cause any trouble about this.

But here I am just assuming, it still would be better to have a confirmation from ST team.
Sébastien Riccio System & Network Admin https://swisscenter.com
1
Webio Replied
Thanks. When I started to check logs for client which reported this issue I've found errors and then this topic so I've enabled for now "Continue delivery if session is disconnected by client" param.
1
echoDreamz Replied
Phil... Welcome to software development... Cant tell you how many times a Windows Update has caused issues, a PHP update a .net update etc. has caused time and pain figuring out what the heck has happened, it was working, now it's not..

I'd suggest moving off the built-in Microsoft SmtpClient class as it is garbage, go to something like Mailkit (Free and open source) or Mailbee (paid).

Reply to Thread