LATEST EXPERIMENTAL BUILD - 5.4-B2014060501

Use this forum if you want to discuss a problem or ask a question related to a hMailServer beta release.

Do you use these builds and find them useful?

YES! Work great! Nice option for the community so please keep them up.
37
79%
YES! Work OK or minor issues can live with. Handy option so keep them up.
2
4%
No I don't but they could be useful so still keep posting them for others.
5
11%
Yes I have but too many issues for me so I'm on the fence. Flip a coin.
1
2%
NO! I'll stick to official releases ONLY so don't bother.
2
4%
 
Total votes: 47

santosh.tarlapally
New user
New user
Posts: 19
Joined: 2013-08-07 07:57

Re: LATEST EXPERIMENTAL BUILD - 5.4-B2013110801

Post by santosh.tarlapally » 2013-12-11 07:20

Thanks for the patches , but the links are from external websites and not from hmailserver. Is this patch provided by hmailserver or is it a 3rd party solution. Please let me know so that i can apply the patch and test . I am struggling with this hmail iocp exception from long time.

Bill48105
Developer
Developer
Posts: 6192
Joined: 2010-04-24 23:16
Location: Michigan, USA

Re: LATEST EXPERIMENTAL BUILD - 5.4-B2013110801

Post by Bill48105 » 2013-12-11 08:17

santosh.tarlapally wrote:Thanks for the patches , but the links are from external websites and not from hmailserver. Is this patch provided by hmailserver or is it a 3rd party solution. Please let me know so that i can apply the patch and test . I am struggling with this hmail iocp exception from long time.
The patches are unofficial & not hosted on hmailserver.com
Bill
hMailServer build LIVE on my servers: 5.4-B2014050402
#hmailserver on FreeNode IRC https://webchat.freenode.net/?channels=#hmailserver
*** ABSENT FROM hMail! Those in IRC know how to find me if urgent. ***

santosh.tarlapally
New user
New user
Posts: 19
Joined: 2013-08-07 07:57

Re: LATEST EXPERIMENTAL BUILD - 5.4-B2013110801

Post by santosh.tarlapally » 2013-12-11 08:28

Ok

can i know when these patches are likely to become official and incorporated in to the hmail release.

Bill48105
Developer
Developer
Posts: 6192
Joined: 2010-04-24 23:16
Location: Michigan, USA

Re: LATEST EXPERIMENTAL BUILD - 5.4-B2013110801

Post by Bill48105 » 2013-12-11 09:28

santosh.tarlapally wrote:Ok

can i know when these patches are likely to become official and incorporated in to the hmail release.
Sorry do not know. Development has pretty much been on hold since the release of 5.4. Overall 5.4 is very stable for most people & we've been gathering feedback on where to go next.

Just so you know myself & others have used my experimental builds in production servers for a couple years now. Generally speaking they are stable except where noted. As a matter of fact I upgraded 2 servers hosting 80+ domains & many 100's of users with the latest experimental & it's been running great. Generally speaking I agree most people should stick with release builds but for those who are having specific problems addressed with the experimental builds it's worth taking a closer look at them.

I'll get with martin & talk about when 5.4.1 might be out but I doubt it'll be any time soon unless we get enough feedback from people that the changes done in the newest experimentals are helpful & don't break anything.
Bill
hMailServer build LIVE on my servers: 5.4-B2014050402
#hmailserver on FreeNode IRC https://webchat.freenode.net/?channels=#hmailserver
*** ABSENT FROM hMail! Those in IRC know how to find me if urgent. ***

Bill48105
Developer
Developer
Posts: 6192
Joined: 2010-04-24 23:16
Location: Michigan, USA

Re: LATEST EXPERIMENTAL BUILD - 5.4-B2013110801

Post by Bill48105 » 2013-12-12 19:24

As an update, I've been running the 11/8/2013 experimental live on 2 servers without any new problems for over a month now. The odd indexing/database errors due to encoded subject lines appear to have stopped that's for sure. I can't say the IOCP problems have been fixed completely but it is believed that update should help with SOME people having IOCP errors, depending on the root cause. It is believed there could be other areas in the code similar to the one found & changed but at least now we have a better idea of what to look for. It's a matter of tracking them down & help from those having IOCP crashes/errors with the latest code would assist greatly. :)
Bill
hMailServer build LIVE on my servers: 5.4-B2014050402
#hmailserver on FreeNode IRC https://webchat.freenode.net/?channels=#hmailserver
*** ABSENT FROM hMail! Those in IRC know how to find me if urgent. ***

marrco
New user
New user
Posts: 6
Joined: 2011-08-04 16:29

Re: LATEST EXPERIMENTAL BUILD - 5.4-B2013110801

Post by marrco » 2013-12-20 12:07

the new build still has the same problem. On a specific server that's multihomed i had hmailserver block, not accepting new connections. On Status / number of open session / stmp is struck to 14

Clicking stop and starting hmail again withing the control panel made hmail accept new connection, but the 14 frozen connections remained. Restarting the server fix the issue.

Problem happened 3 times in the last few months and at least 2 times i also had problems with one internet connections, going up and down. So my educated guess it that some stmp transaction (or maybe virus scan processes) got truncated because of one problematic network path and never got cleared.

Bill48105
Developer
Developer
Posts: 6192
Joined: 2010-04-24 23:16
Location: Michigan, USA

Re: LATEST EXPERIMENTAL BUILD - 5.4-B2013110801

Post by Bill48105 » 2013-12-20 22:14

marrco wrote:the new build still has the same problem. On a specific server that's multihomed i had hmailserver block, not accepting new connections. On Status / number of open session / stmp is struck to 14

Clicking stop and starting hmail again withing the control panel made hmail accept new connection, but the 14 frozen connections remained. Restarting the server fix the issue.

Problem happened 3 times in the last few months and at least 2 times i also had problems with one internet connections, going up and down. So my educated guess it that some stmp transaction (or maybe virus scan processes) got truncated because of one problematic network path and never got cleared.
"same problem" is unlikely but know what you meant. The issue is there are many places causing the iocp crashes/hangs it's just a matter of tracking them all down & where (if) they happen depends on each server. There is no rhyme or reason really as to why some people have the problem & others don't. I run 2 hmailservers myself that are quite busy & I never see iocp crashes or hangs. Heck I usually can't make it happen if it TRY, even on test servers. So there are unknown variables at play which is likely combination of server hardware/specs, Windows settings, what's installed in Windows, particular hmail settings, usage patterns, etc. The only thing we really know for sure is it's timing related & obviously the timing will vary from server to server & time of day, day of week, etc.

But yes if you're having line problems or if the remote is behaving badly (like spammers) it is very possible you are more prone to have problems. Perhaps one of the reasons I don't have problems on my busier server is that I run ASSP in front of it but the other does not have ASSP & never crashes/hangs. Heck they'll go months & months without restarting hmail or windows unless I'm updating hmail to new version which hasn't been too often lately.

Not sure if it'll help you but someone else claims adjusting async task threads helped them. In your case you might adjust the SMTP timeout INI setting.
Bill
hMailServer build LIVE on my servers: 5.4-B2014050402
#hmailserver on FreeNode IRC https://webchat.freenode.net/?channels=#hmailserver
*** ABSENT FROM hMail! Those in IRC know how to find me if urgent. ***

Bill48105
Developer
Developer
Posts: 6192
Joined: 2010-04-24 23:16
Location: Michigan, USA

Re: LATEST EXPERIMENTAL BUILD - 5.4-B2013110801

Post by Bill48105 » 2013-12-27 23:04

FYI for anyone following this thread, I have an experimental build of hmail available that adds STARTTLS to SMTP port 587 (just for now) if anyone wants to test it on a TEST server/computer (like personal workstation or VM). I'll be posting it up for everyone once we get some basic testing completed. A LOT had to change to accommodate STARTTLS but so far so good in testing done here. Best way to request a copy is to visit us in IRC:
http://www.hmailserver.com/irc_fullscreen.php
Or using IRC client irc.hellscream.org.uk #hmailserver
You can PM me on here is well but I don't always get those for a few days.
Thanks,
Bill
hMailServer build LIVE on my servers: 5.4-B2014050402
#hmailserver on FreeNode IRC https://webchat.freenode.net/?channels=#hmailserver
*** ABSENT FROM hMail! Those in IRC know how to find me if urgent. ***

User avatar
mattg
Moderator
Moderator
Posts: 22435
Joined: 2007-06-14 05:12
Location: 'The Outback' Australia

Re: LATEST EXPERIMENTAL BUILD - 5.4-B2013110801

Post by mattg » 2013-12-28 03:23

Great work Bill.

I know that you've covered this before, but how is this to a) install, and b) remove if things go pear shaped
Just 'cause I link to a page and say little else doesn't mean I am not being nice.
https://www.hmailserver.com/documentation

Bill48105
Developer
Developer
Posts: 6192
Joined: 2010-04-24 23:16
Location: Michigan, USA

Re: LATEST EXPERIMENTAL BUILD - 5.4-B2013110801

Post by Bill48105 » 2013-12-28 06:44

mattg wrote:Great work Bill.

I know that you've covered this before, but how is this to a) install, and b) remove if things go pear shaped
Thx matt. Rolaids0 & I worked on it the past week.

To install you need working hmail 5.4, usually the most recent release as noted so right now it'd be B1950. Once that was installed or upgraded to & working you just backup (especialyl BIN folder but really should backup everything just in case), stop hmail service, drop the files in the 7z file into the bin folder replacing existing then start up hmail again. Then test. To undo you do the reverse. Stop & restore your BIN folder files you replaced & start hmail again.

In the case of this test build the backup step isn't really needed as I WOULD NOT recommend you use it on a production server at this point so I'd hope it was being testing on a TEST BED server. ;)

I can PM you a link to the download. I plan to post it like the other builds but wanted to wait until we did some testing. Once we're sure it's working & user settings are in it'll get committed to the official source for martin to review & hopefully end up in the next release. At that point it'd just be a matter of upgrading like any other hmail version. My builds are just meant to be used between releases & for people who need them like hotfixes or testing new features.
Bill
hMailServer build LIVE on my servers: 5.4-B2014050402
#hmailserver on FreeNode IRC https://webchat.freenode.net/?channels=#hmailserver
*** ABSENT FROM hMail! Those in IRC know how to find me if urgent. ***

markusr
New user
New user
Posts: 1
Joined: 2014-01-04 14:22

Re: LATEST EXPERIMENTAL BUILD - 5.4-B2013110801

Post by markusr » 2014-01-04 14:35

Hello,

i´ve set up hmailserver and some other mailserver for testing. At the end i would like to have a runing w2kserver (later, mybe this year :) ) with webmail and so on.
If i could help please give me a link to this latest build of hmailserver. I´m very interested in working with this version, thank you.


regards markus

(english is not my native language, so sorry for any mistakes in writing)

Bill48105
Developer
Developer
Posts: 6192
Joined: 2010-04-24 23:16
Location: Michigan, USA

Re: LATEST EXPERIMENTAL BUILD - 5.4-B2013110801

Post by Bill48105 » 2014-01-04 18:57

markusr wrote:Hello,

i´ve set up hmailserver and some other mailserver for testing. At the end i would like to have a runing w2kserver (later, mybe this year :) ) with webmail and so on.
If i could help please give me a link to this latest build of hmailserver. I´m very interested in working with this version, thank you.


regards markus

(english is not my native language, so sorry for any mistakes in writing)
Hi.
I can send you PM markusr. Click on Messages link to view.
Bill
hMailServer build LIVE on my servers: 5.4-B2014050402
#hmailserver on FreeNode IRC https://webchat.freenode.net/?channels=#hmailserver
*** ABSENT FROM hMail! Those in IRC know how to find me if urgent. ***

Bill48105
Developer
Developer
Posts: 6192
Joined: 2010-04-24 23:16
Location: Michigan, USA

Re: LATEST EXPERIMENTAL BUILD - 5.4-B2013110801

Post by Bill48105 » 2014-01-18 22:13

ANYONE WHO USES HMAIL'S PHPWEBADMIN SHOULD SEE THIS IMMEDIATELY http://www.hmailserver.com/forum/viewto ... 86#p157086
hMailServer build LIVE on my servers: 5.4-B2014050402
#hmailserver on FreeNode IRC https://webchat.freenode.net/?channels=#hmailserver
*** ABSENT FROM hMail! Those in IRC know how to find me if urgent. ***

ShiningDragon
New user
New user
Posts: 6
Joined: 2011-08-12 13:58

Re: LATEST EXPERIMENTAL BUILD - 5.4-B2013110801

Post by ShiningDragon » 2014-01-19 05:05

Hi!

Hopefully i'm within the right thread.

I've upgrade to the latest experimental build (5.4-B2013110801) and are not able anymore to use PHPWebAdmin anymore on my server.

Everytime i try to access this one, i'm getting this:
Operation failed
The operation failed. Please make sure that you have logged on with the appropriate permissions to perform this task.

The following description exists:

include(): 1. h->opened_path=[null] h->filename=[hm_login.php]

Error location: index.php
It worked before the upgrading. The versioncheck was updated with the versionnumber of the experimentalbuild.

I am running hMailServer on a Windows 7 x86-64 machine with Apache and PHP enabled. I've readed all PHPWebAdmin related FAQs on this site (setting up, error handling and some threads) but couldn't find this problem.
I've followed the post from Bill and added the .php and changed the background-login.php, just to be sure that the last upgrade isn't missing some of the code.

This is the login for the WebAdmin and here's my phpinfo of the current installation.

Any tips and hints or known php compability-problems for this? ^^;
I love hMailServer - my one and only mail-solution.

Bill48105
Developer
Developer
Posts: 6192
Joined: 2010-04-24 23:16
Location: Michigan, USA

Re: LATEST EXPERIMENTAL BUILD - 5.4-B2013110801

Post by Bill48105 » 2014-01-19 05:52

ShiningDragon wrote:Hi!

Hopefully i'm within the right thread.

I've upgrade to the latest experimental build (5.4-B2013110801) and are not able anymore to use PHPWebAdmin anymore on my server.

Everytime i try to access this one, i'm getting this:
Operation failed
The operation failed. Please make sure that you have logged on with the appropriate permissions to perform this task.

The following description exists:

include(): 1. h->opened_path=[null] h->filename=[hm_login.php]

Error location: index.php
It worked before the upgrading. The versioncheck was updated with the versionnumber of the experimentalbuild.

I am running hMailServer on a Windows 7 x86-64 machine with Apache and PHP enabled. I've readed all PHPWebAdmin related FAQs on this site (setting up, error handling and some threads) but couldn't find this problem.
I've followed the post from Bill and added the .php and changed the background-login.php, just to be sure that the last upgrade isn't missing some of the code.

This is the login for the WebAdmin and here's my phpinfo of the current installation.

Any tips and hints or known php compability-problems for this? ^^;
Hello. This thread really is just for experimental-specific issues & discussion. That is definitely an odd error & more so that the problem would start after upgrading. Did you start fresh with new phpwebadmin from 5.4 B1950 or did you re-use some from old admin? You might start fresh just to test.

If you have time it might be easier if you stop into IRC & chat about it.
Thx
Bill
hMailServer build LIVE on my servers: 5.4-B2014050402
#hmailserver on FreeNode IRC https://webchat.freenode.net/?channels=#hmailserver
*** ABSENT FROM hMail! Those in IRC know how to find me if urgent. ***

Bill48105
Developer
Developer
Posts: 6192
Joined: 2010-04-24 23:16
Location: Michigan, USA

Re: LATEST EXPERIMENTAL BUILD - 5.4-B2014010101

Post by Bill48105 » 2014-01-25 09:34

NEW BUILDS POSTED. LOTS OF CHANGES!!!
2014-01-25 5.4-B2014012501
* IMPORTANT: This build has a LOT of extra debug logging by default. Disable using [Settings]LogLevel=7 or lower
* Added multi-forwards so now "Forward" on a user can be more than 1 address using commas. GUI has not been updated so for now setup like: [user]@[domain.dom,anotheruser@domain.dom] IOW just add the rest in the domain field for now
* Added multi-aliases so now "Aliases" can be more than 1 address using commas. GUI was updated previous thanks to Rolaids0 so just enter like: [user@domain.dom,anotheruser@domain.dom]

2014-01-15 TEST-15Jan14-ALPHA
* IMPORTANT: This build has a LOT of extra debug logging by default. Disable using [Settings]LogLevel=7 or lower
* STARTTLS SMTP incoming only. Setup on Ports like SSL. Check Use STARTTLS (Thanks Rolaids0 for helping)
* SPF default (if sender has no SPF) & override (to alter sender policy such as don't allow +all) policies using INI settings
* OnSMTPData now has oClient.STARTTLS such as If (oClient.Port = "587" And (oClient.Username = "" Or oClient.STARTTLS = "")) Then Result.Value=1 'Reject it
* ESMTPx headers to show if sender was AUTH'd or used STARTTLS such as ESMTPA ESMTPS or ESMTPSA for both
* Negative DNS blacklists now possible allowing for DNS "whitelists" such as dnswl.org setup just like blacklist but set negative score to subtract (Thanks to Rolaids0 for helping)
* Valid email address pattern ini allows over-ride of what hmail thinks is valid. Helpful for @localhost @fax or workflow@noreply for MS Sharepoint (no suffix)
* Fix for UTF-8 indexing "DALConnection::Execute, Description: MySQL: Incorrect string value" errors (Thanks greylock!)
* Protocol parser critical section added to stop IOCP crashes (Thanks greylock!)
* Tcpconnection critical sections added to stop possible email corruption & IOCP crashes
* SMTP delivery manager extra iocp error logging added
* Auto responder "FROM" can now be set in ini so can set <> or noreply@yourdomain or mailer-daemon@ etc rather than hmail using recipient's address
* Added extra auto responder logging
* Added fired event logging so now can see which events were fired in logs (Thanks Rolaids0!)
* Added messagesize debug logging
* BE SURE TO SEE NEW INI's IN POST BELOW
hMailServer build LIVE on my servers: 5.4-B2014050402
#hmailserver on FreeNode IRC https://webchat.freenode.net/?channels=#hmailserver
*** ABSENT FROM hMail! Those in IRC know how to find me if urgent. ***

User avatar
mattg
Moderator
Moderator
Posts: 22435
Joined: 2007-06-14 05:12
Location: 'The Outback' Australia

Re: LATEST EXPERIMENTAL BUILD - 5.4-B2014010101

Post by mattg » 2014-01-25 14:06

Are you using this on a production machine Bill?
Should we?
Just 'cause I link to a page and say little else doesn't mean I am not being nice.
https://www.hmailserver.com/documentation

Bill48105
Developer
Developer
Posts: 6192
Joined: 2010-04-24 23:16
Location: Michigan, USA

Re: LATEST EXPERIMENTAL BUILD - 5.4-B2014010101

Post by Bill48105 » 2014-01-25 17:47

mattg wrote:Are you using this on a production machine Bill?
Should we?
I've been using the Jan 15th ALPHA on a production server for 10 days now without a problem so should be quite stable. Granted I'm not using STARTTLS yet but am using SSL & all the other new settings. Unfortunately each server is different so hard to say but it SHOULD be OK. :) The 25th build had some pretty major changes done in forwarding/aliases & I've not used that on my live server yet but testing I've done shows it should be fine especially if you don't make use of multi-fowards or multi-aliases because it SHOULD act the same without them.

A few people have tested the 15th build or other STARTTLS-era builds & no reported problems but in the end we won't know for sure until people try em out. :)
Bill
hMailServer build LIVE on my servers: 5.4-B2014050402
#hmailserver on FreeNode IRC https://webchat.freenode.net/?channels=#hmailserver
*** ABSENT FROM hMail! Those in IRC know how to find me if urgent. ***

prisma
Senior user
Senior user
Posts: 325
Joined: 2010-07-09 13:16

Re: LATEST EXPERIMENTAL BUILD - 5.4-B2014010101

Post by prisma » 2014-01-27 11:30

Found a possible Memory bug in hMailadmin. I'm using the latest experimental build. This error is shown on Windows 8 64bit when clicking on settings/advanced/smtp. HmailAdmin is started via "\\myServer\c$\Program Files\hMailserver\Bin". When started directly on the Server 2008 32bit server via "C:\Program Files\hMailserver\Bin" this error does NOT appear:

Code: Select all

ExceptionType: AccessViolationException
HelpLine: 
Message:Tried to read in or write protected memory. This is often an indication that other memory is corrupt. (Translation by translate.google.com)
Source: Interop.hMailServer
StackTrace:    bei hMailServer.TCPIPPortClass.get_UseSTARTTLS()
   bei hMailServer.Administrator.ucTCPIPPort.LoadData()
   bei hMailServer.Administrator.formMain.ShowNodeRepresentation(INode node)
   bei hMailServer.Administrator.formMain.treeNodes_AfterSelect(Object sender, TreeViewEventArgs e)
   bei System.Windows.Forms.TreeView.OnAfterSelect(TreeViewEventArgs e)
   bei System.Windows.Forms.TreeView.TvnSelected(NMTREEVIEW* nmtv)
   bei System.Windows.Forms.TreeView.WmNotify(Message& m)
   bei System.Windows.Forms.TreeView.WndProc(Message& m)
   bei System.Windows.Forms.Control.ControlNativeWindow.OnMessage(Message& m)
   bei System.Windows.Forms.Control.ControlNativeWindow.WndProc(Message& m)
   bei System.Windows.Forms.NativeWindow.Callback(IntPtr hWnd, Int32 msg, IntPtr wparam, IntPtr lparam)
TargetSite: Boolean get_UseSTARTTLS()

Bill48105
Developer
Developer
Posts: 6192
Joined: 2010-04-24 23:16
Location: Michigan, USA

Re: LATEST EXPERIMENTAL BUILD - 5.4-B2014010101

Post by Bill48105 » 2014-01-27 18:09

prisma wrote:Found a possible Memory bug in hMailadmin. I'm using the latest experimental build. This error is shown on Windows 8 64bit when clicking on settings/advanced/smtp. HmailAdmin is started via "\\myServer\c$\Program Files\hMailserver\Bin". When started directly on the Server 2008 32bit server via "C:\Program Files\hMailserver\Bin" this error does NOT appear:

Code: Select all

ExceptionType: AccessViolationException
HelpLine: 
Message:Tried to read in or write protected memory. This is often an indication that other memory is corrupt. (Translation by translate.google.com)
Source: Interop.hMailServer
StackTrace:    bei hMailServer.TCPIPPortClass.get_UseSTARTTLS()
   bei hMailServer.Administrator.ucTCPIPPort.LoadData()
   bei hMailServer.Administrator.formMain.ShowNodeRepresentation(INode node)
   bei hMailServer.Administrator.formMain.treeNodes_AfterSelect(Object sender, TreeViewEventArgs e)
   bei System.Windows.Forms.TreeView.OnAfterSelect(TreeViewEventArgs e)
   bei System.Windows.Forms.TreeView.TvnSelected(NMTREEVIEW* nmtv)
   bei System.Windows.Forms.TreeView.WmNotify(Message& m)
   bei System.Windows.Forms.TreeView.WndProc(Message& m)
   bei System.Windows.Forms.Control.ControlNativeWindow.OnMessage(Message& m)
   bei System.Windows.Forms.Control.ControlNativeWindow.WndProc(Message& m)
   bei System.Windows.Forms.NativeWindow.Callback(IntPtr hWnd, Int32 msg, IntPtr wparam, IntPtr lparam)
TargetSite: Boolean get_UseSTARTTLS()
Hey Prisma. Thanks for the info. Not a memory bug it is related to changes in hmail's COM system. It happens when the admin & server our "out of sync" version-wise. Are you SURE you replaced ALL FILES in the bin folder with ALL FILES in the 7z? (Granted there are fewer files in the 7z. No INI for example as you would never want to replace that but you do need to replace all that do exist in the 7z including dll's & the tlb file. (The dll's & tlb are what allow the admin & server to talk to each other via COM)

In rare cases you might need to register the changes but that should not normally be needed at least I've never had to do it on my server or test servers. In your case it's likely how you are running it over the network. Did you update all files on both computers? (I've never run it like that myself but if it worked before it should work after changes as long as you updated all the files)
Bill
hMailServer build LIVE on my servers: 5.4-B2014050402
#hmailserver on FreeNode IRC https://webchat.freenode.net/?channels=#hmailserver
*** ABSENT FROM hMail! Those in IRC know how to find me if urgent. ***

prisma
Senior user
Senior user
Posts: 325
Joined: 2010-07-09 13:16

Re: LATEST EXPERIMENTAL BUILD - 5.4-B2014010101

Post by prisma » 2014-01-27 18:43

On the 64bit Win8 are only the admin components installed. and no, they are completely un-updated, only the server components are from the alpha. But I started the updated components over the UNC-path. It was not clear for me, that the local components can influence the scenario. sorry for my confusion.

Bill48105
Developer
Developer
Posts: 6192
Joined: 2010-04-24 23:16
Location: Michigan, USA

Re: LATEST EXPERIMENTAL BUILD - 5.4-B2014010101

Post by Bill48105 » 2014-01-27 19:06

prisma wrote:On the 64bit Win8 are only the admin components installed. and no, they are completely un-updated, only the server components are from the alpha. But I started the updated components over the UNC-path. It was not clear for me, that the local components can influence the scenario. sorry for my confusion.
Yes you should replace the files on the admin-only computer too. It needs the right interop & common dll's WHERE IT WAS INSTALLLED even if you run over the network. Or you might need to copy the files from the 7z to the admin-only bin & try. If that doesn't work you might need to run one or both of the register commands:

Code: Select all

hmailserver.exe /register
hmailserver.exe /registertlb
Not sure if reboot is needed or not but unlikely. Know that I have NOT had to run those on any test I've done as long as I replaced all the exe/dll/tlb files in the bin folder with hmail service stopped & all hmail GUI's closed. So running those isn't likely needed but in your case running over LAN maybe it is as I've not tried that.
Bill
hMailServer build LIVE on my servers: 5.4-B2014050402
#hmailserver on FreeNode IRC https://webchat.freenode.net/?channels=#hmailserver
*** ABSENT FROM hMail! Those in IRC know how to find me if urgent. ***

Bill48105
Developer
Developer
Posts: 6192
Joined: 2010-04-24 23:16
Location: Michigan, USA

Re: LATEST EXPERIMENTAL BUILD - 5.4-B2014020501

Post by Bill48105 » 2014-02-05 09:09

New build posted:

2014-02-05 5.4-B2014020501
* IMPORTANT: This build has a LOT of extra debug logging by default. Disable using [Settings]LogLevel=7 or lower
* Adjusted Received headers so when AuthUserReplacementIP is set that HELO is localhost or Unknown
* http://www.hmailserver.com/forum/viewto ... 07#p157707


Also updated INI examples to include AddXAuthUserIP

FYI: So far so good with Jan 15th build based on reports from numerous people testing it & using it live.
hMailServer build LIVE on my servers: 5.4-B2014050402
#hmailserver on FreeNode IRC https://webchat.freenode.net/?channels=#hmailserver
*** ABSENT FROM hMail! Those in IRC know how to find me if urgent. ***

Kriztan
Normal user
Normal user
Posts: 39
Joined: 2010-03-17 16:12
Location: Germany

Re: LATEST EXPERIMENTAL BUILD - 5.4-B2014020501

Post by Kriztan » 2014-02-05 10:12

Hi,
I'm using the LATEST EXPERIMENTAL BUILD - 5.4-B2014020501 on my system. Is it possible to use STARTTLS for relaying mails to a smart relay on port 25? My hoster only supports plain text auth and STARTTLS on port 25 together.
Thanks Kriztan

Bill48105
Developer
Developer
Posts: 6192
Joined: 2010-04-24 23:16
Location: Michigan, USA

Re: LATEST EXPERIMENTAL BUILD - 5.4-B2014020501

Post by Bill48105 » 2014-02-05 10:34

Kriztan wrote:Hi,
I'm using the LATEST EXPERIMENTAL BUILD - 5.4-B2014020501 on my system. Is it possible to use STARTTLS for relaying mails to a smart relay on port 25? My hoster only supports plain text auth and STARTTLS on port 25 together.
Thanks Kriztan
Hi Kriztan,
No sorry at this time hmail only supports STARTTLS for INCOMING SMTP not OUTGOING SMTP. Are you sure they don't support SSL/TLS vs STARTTLS? Maybe on port 465?

Btw feel free to vote/comment about what's next for STARTTLS in hmail here:
http://www.hmailserver.com/forum/viewto ... =2&t=25966
Thx
Bill
hMailServer build LIVE on my servers: 5.4-B2014050402
#hmailserver on FreeNode IRC https://webchat.freenode.net/?channels=#hmailserver
*** ABSENT FROM hMail! Those in IRC know how to find me if urgent. ***

Kriztan
Normal user
Normal user
Posts: 39
Joined: 2010-03-17 16:12
Location: Germany

Re: LATEST EXPERIMENTAL BUILD - 5.4-B2014020501

Post by Kriztan » 2014-02-05 11:03

Bill48105 wrote:Are you sure they don't support SSL/TLS vs STARTTLS? Maybe on port 465?
Hi Bill,

I've already asked my hoster, but this is a time ago. He said, that SSL is only supported via STARTTLS. The classic SSL is not supported and SSL via port 465 was not working at this time.
I'll test it now with this release --> working with checked SSL box and port 465!

I've voted for STARTTLS for outgoing connections ;)

Thanks Kriztan

Kriztan
Normal user
Normal user
Posts: 39
Joined: 2010-03-17 16:12
Location: Germany

Re: LATEST EXPERIMENTAL BUILD - 5.4-B2014020501

Post by Kriztan » 2014-02-05 14:37

Hil Bill,
I've found a strange behaviour:
When I send a mail with an local account with the hmail built-in signature, the mail message-area is empty. After deactivating the built in signature the mails arrive completely.
Can you check this?
Thanks Kriztan

prisma
Senior user
Senior user
Posts: 325
Joined: 2010-07-09 13:16

Re: LATEST EXPERIMENTAL BUILD - 5.4-B2014020501

Post by prisma » 2014-02-07 15:19

+1 outgoing SMTP.

I assume the most important thing is to start with STARTTLS for "SMTP Relayer" and routes with "Target SMTP host". It's true, some European providers cut of unencrypted delivery and only provide enforced STARTTLS on 587, no SSL/TLS on 465. This fact is intensified by the fact that in Europe the use of a "Smart Host" (to use M$ diction) is often mandatory, even if you have a static IP address.

At the moment there's no visible switch, only a script hook, to configure enforcement of STARTTLS for incoming SMTP. I think it could make sense to have at least a enforcement hook for Relayer and Routes also. Of course, better was a visible switch for enforcement for in and out... but... I already discussed this with Bill.

Second step could be implementing optional STARTTLS (and always optional to do not break RFC) for delivery on 25 to resolved MX records. (Look if ESMTPS/STARTTLS is provided, if not deliver unencrypted).

I think, at least the second step is what the Germans would prefer ;) LOL ... I'm still hopefully.

Bill48105
Developer
Developer
Posts: 6192
Joined: 2010-04-24 23:16
Location: Michigan, USA

Re: LATEST EXPERIMENTAL BUILD - 5.4-B2014020501

Post by Bill48105 » 2014-02-07 16:40

I hear ya prisma no need to reply on both posts :P

That's unfortunate Kriztan. Not sure why an ISP could support STARTTLS & not SSL unless they were just lazy. And as far as your signatures problem, not sure on that. Doubt anything was changed in experimental that would cause that. Please post up a new thread & tell the version you are using.
Thx
hMailServer build LIVE on my servers: 5.4-B2014050402
#hmailserver on FreeNode IRC https://webchat.freenode.net/?channels=#hmailserver
*** ABSENT FROM hMail! Those in IRC know how to find me if urgent. ***

marrco
New user
New user
Posts: 6
Joined: 2011-08-04 16:29

Re: LATEST EXPERIMENTAL BUILD - 5.4-B2014020501

Post by marrco » 2014-02-07 17:57

I had an other crash a few days ago, and now i updated to the new 2014/01/25 version.

This is in the error log:

"ERROR" 1824 "2014-02-04 13:35:00.449" "Severity: 2 (High), Code: HM4208, Source: IOCPQueueWorkerTask::DoWork, Description: An unknown error occured while handling asynchronous requests. Error number: invalid vector<T> subscript"

hth
marco

Bill48105
Developer
Developer
Posts: 6192
Joined: 2010-04-24 23:16
Location: Michigan, USA

Re: LATEST EXPERIMENTAL BUILD - 5.4-B2014020501

Post by Bill48105 » 2014-02-07 20:12

marrco wrote:I had an other crash a few days ago, and now i updated to the new 2014/01/25 version.

This is in the error log:

"ERROR" 1824 "2014-02-04 13:35:00.449" "Severity: 2 (High), Code: HM4208, Source: IOCPQueueWorkerTask::DoWork, Description: An unknown error occured while handling asynchronous requests. Error number: invalid vector<T> subscript"

hth
marco
Hey marco,
Were you having crashes before the new build? Did hmail crash or just log that error? If you can PM me more log lines around that error I'd appreciate it.
Thx
Bill
hMailServer build LIVE on my servers: 5.4-B2014050402
#hmailserver on FreeNode IRC https://webchat.freenode.net/?channels=#hmailserver
*** ABSENT FROM hMail! Those in IRC know how to find me if urgent. ***

Bill48105
Developer
Developer
Posts: 6192
Joined: 2010-04-24 23:16
Location: Michigan, USA

Re: LATEST EXPERIMENTAL BUILD - 5.4-B2014020501

Post by Bill48105 » 2014-02-08 04:04

NEW TEST BUILD POSTED:

2014-02-07 5.4-B2014020701
* IMPORTANT: This build has a LOT of extra debug logging by default. Disable using [Settings]LogLevel=7 or lower
* SPECIAL TEST BUILD - Addresses generated file names in IMAP FETCH that are not RFC compliant
* http://www.hmailserver.com/forum/viewto ... =7&t=25985
hMailServer build LIVE on my servers: 5.4-B2014050402
#hmailserver on FreeNode IRC https://webchat.freenode.net/?channels=#hmailserver
*** ABSENT FROM hMail! Those in IRC know how to find me if urgent. ***

Bill48105
Developer
Developer
Posts: 6192
Joined: 2010-04-24 23:16
Location: Michigan, USA

Re: LATEST EXPERIMENTAL BUILD - 5.4-B2014020801

Post by Bill48105 » 2014-02-08 08:03

NEW TEST BUILD:

2014-02-08 5.4-B2014020801
* IMPORTANT: This build has a LOT of extra debug logging by default. Disable using [Settings]LogLevel=7 or lower
* Added ability to set custom DNS servers in INI DnsServers=192.168.1.33
* Only 1 DNS server for now! IPv6 compatibility UNKNOWN
hMailServer build LIVE on my servers: 5.4-B2014050402
#hmailserver on FreeNode IRC https://webchat.freenode.net/?channels=#hmailserver
*** ABSENT FROM hMail! Those in IRC know how to find me if urgent. ***

User avatar
mattg
Moderator
Moderator
Posts: 22435
Joined: 2007-06-14 05:12
Location: 'The Outback' Australia

Re: LATEST EXPERIMENTAL BUILD - 5.4-B2014010101

Post by mattg » 2014-02-10 01:50

Bill48105 wrote:
mattg wrote:Are you using this on a production machine Bill?
Should we?
I've been using the Jan 15th ALPHA on a production server for 10 days now without a problem so should be quite stable. Granted I'm not using STARTTLS yet but am using SSL & all the other new settings. Unfortunately each server is different so hard to say but it SHOULD be OK. :) The 25th build had some pretty major changes done in forwarding/aliases & I've not used that on my live server yet but testing I've done shows it should be fine especially if you don't make use of multi-fowards or multi-aliases because it SHOULD act the same without them.

A few people have tested the 15th build or other STARTTLS-era builds & no reported problems but in the end we won't know for sure until people try em out. :)
Bill
OK Bill

I have put this version (15th Jan) on my production server...

How can I tell if the log level setting that I set is taking effect?
The [settings] section of the INI should be logged on restart I'd reckon....

I have put LogLevel=6 but the debug logging is still a bit full on. I don't need to know every TCPConnection::Write SSL or without SSL It seems to be writing these debug lines for all IMAP messages - I don't normally log IMAP connections unless something isn't working... (again too verbose for my liking)
I like debug logging normally but that statement alone just the log size tenfold in less than 10 minutes. I can see that it would be great for testing, but not at all times.

LogLevel=4 still has these lines
LogLevel=1 seems to have fixed them
Just 'cause I link to a page and say little else doesn't mean I am not being nice.
https://www.hmailserver.com/documentation

Bill48105
Developer
Developer
Posts: 6192
Joined: 2010-04-24 23:16
Location: Michigan, USA

Re: LATEST EXPERIMENTAL BUILD - 5.4-B2014010101

Post by Bill48105 » 2014-02-10 03:40

mattg wrote: OK Bill

I have put this version (15th Jan) on my production server...

How can I tell if the log level setting that I set is taking effect?
The [settings] section of the INI should be logged on restart I'd reckon....

I have put LogLevel=6 but the debug logging is still a bit full on. I don't need to know every TCPConnection::Write SSL or without SSL It seems to be writing these debug lines for all IMAP messages - I don't normally log IMAP connections unless something isn't working... (again too verbose for my liking)
I like debug logging normally but that statement alone just the log size tenfold in less than 10 minutes. I can see that it would be great for testing, but not at all times.

LogLevel=4 still has these lines
LogLevel=1 seems to have fixed them
Hey Matt. The Loglevel stuff is pretty much undocumented. In essence when I'm adding debug logging I SOMETIMES check LogLevel setting. What that means is SOMETIMES extra debug lines will show up if Debug logging is enabled & there is no way to turn off & sometimes those lines will only show up if LogLevel is high enough. The level needed is arbitrary in my head at the time I'm doing the changes kind of guessing how important the log lines are to show or not when debug logging is enabled. For very annoying very frequent logging (such as socket read/writes) I'd check that level is at least 9 likely as very unlikely someone would want that especially on a busy server but I want a way for someone to turn it on if needed thus 9 or the highest. Some things I'd set at like 7 where others could be 5 3 or 1 depending how annoying or resource intensive the logging would be. The problem like I said is I never posted what log levels are needed for which lines. I suppose I should have the log line itself state the level then it'd be clear what level was needed to not have it logged. And yeah suppose it might be nice to log some settings upon startup.

Btw the loglevel stuff is a fallback to the old MTS Pro mail server I used (and helped develop) which had that & we found it very useful even if slightly confusing. The idea is there should be some granular logging output instead of just on or off. The issue is someone deciding what level should be req'd for each logging line & at this point it's just me guessing what would be reasonable. If there is a particular log line that is troublesome to you (realize I use these test builds on my live servers so generally I'm pretty sensitive to annoyances myself) then please let me know & I'll make an adjustment & post a new build. If all else fails set LogLevel=0 or 1 or turn off debug logging.
Bill

PS. I went to grab an example & think I know of one that is annoying that for some reason I set to 2:

Code: Select all

if (IniFileSettings::Instance()->GetLogLevel() > 2) LOG_DEBUG("TCPCconnection::GetSocket(): next_layer()");

if (IniFileSettings::Instance()->GetLogLevel() > 2) LOG_DEBUG("TCPCconnection::GetSocket(): lowest_layer()");
For now set LogLevel to 2 or less to not have that & the first_layer lines show but I've changed it to >8 in the code. (So it'd only show for 9+) I need to do some code cleanup anyway so I can merge my changes with the official code & I'll review some of the loglevel's need to be adjusted or even added to avoid unnecessary debug logging. Sometimes I toss in logging when testing here & forget to go back & remove or add loglevel. Since I keep LogLevel=1 on my servers I rarely run into seeing them,

PPS. OK so I went thru the code & quite a few places the loglevels were lower than I thought they were. What i've decided to do is since 9 is the default anything 10+ is 'super debug logging'. I went thru & adjusted all the ones with loglevel & most have been changed to >9. The exception are the loglevel lines I added to alter default behavior of pop & IMAP so I set those to >8 so they show up by default as they did before. Going forward LogLevel will need to be 10 or higher for the extra debug logging to show up unless there is a specific need to do otherwise. I'll get a new build posted up but if you are more comfortable using the Jan 15th build just set loglevel to <2 & you shouldn't see much more logging than you normally would with debug enabled.
hMailServer build LIVE on my servers: 5.4-B2014050402
#hmailserver on FreeNode IRC https://webchat.freenode.net/?channels=#hmailserver
*** ABSENT FROM hMail! Those in IRC know how to find me if urgent. ***

User avatar
mattg
Moderator
Moderator
Posts: 22435
Joined: 2007-06-14 05:12
Location: 'The Outback' Australia

Re: LATEST EXPERIMENTAL BUILD - 5.4-B2014020801

Post by mattg » 2014-02-10 04:28

I'm happy enough at the moment.

I'll watch for some more builds though. SO far so good.
Just 'cause I link to a page and say little else doesn't mean I am not being nice.
https://www.hmailserver.com/documentation

Bill48105
Developer
Developer
Posts: 6192
Joined: 2010-04-24 23:16
Location: Michigan, USA

Re: LATEST EXPERIMENTAL BUILD - 5.4-B2014020801

Post by Bill48105 » 2014-02-10 04:43

mattg wrote:I'm happy enough at the moment.

I'll watch for some more builds though. SO far so good.
ok cool beans :)
hMailServer build LIVE on my servers: 5.4-B2014050402
#hmailserver on FreeNode IRC https://webchat.freenode.net/?channels=#hmailserver
*** ABSENT FROM hMail! Those in IRC know how to find me if urgent. ***

User avatar
mattg
Moderator
Moderator
Posts: 22435
Joined: 2007-06-14 05:12
Location: 'The Outback' Australia

Re: LATEST EXPERIMENTAL BUILD - 5.4-B2014010101

Post by mattg » 2014-02-10 07:08

Bill48105 wrote:NEW BUILDS POSTED. LOTS OF CHANGES!!!
* SPF default (if sender has no SPF) & override (to alter sender policy such as don't allow +all) policies using INI settings
G'day Bill,

Not really sure how to implement this one...

Does simply un-commenting first line override +all SPF records?
that's what I want to achieve - replace all +all with a -all
What about ~all? are they affected

What if someone validly has a specific IP (like say a bulk mailer) allowed under SPF to send, but that isn't their MX or A record?

;SPFPolicyOverride=v=spf1 -all
; Set SPF policy to override sender's policy such as block +all
; Default is do not override sender's policy


Thanks BTW.

Matt
Just 'cause I link to a page and say little else doesn't mean I am not being nice.
https://www.hmailserver.com/documentation

Bill48105
Developer
Developer
Posts: 6192
Joined: 2010-04-24 23:16
Location: Michigan, USA

Re: LATEST EXPERIMENTAL BUILD - 5.4-B2014010101

Post by Bill48105 » 2014-02-10 08:26

mattg wrote:
Bill48105 wrote:NEW BUILDS POSTED. LOTS OF CHANGES!!!
* SPF default (if sender has no SPF) & override (to alter sender policy such as don't allow +all) policies using INI settings
G'day Bill,

Not really sure how to implement this one...

Does simply un-commenting first line override +all SPF records?
that's what I want to achieve - replace all +all with a -all
What about ~all? are they affected

What if someone validly has a specific IP (like say a bulk mailer) allowed under SPF to send, but that isn't their MX or A record?

;SPFPolicyOverride=v=spf1 -all
; Set SPF policy to override sender's policy such as block +all
; Default is do not override sender's policy


Thanks BTW.

Matt
Howdy Matt.

If you set this:
SPFPolicyOverride=v=spf1 -all
It SHOULDl replace whatever they have for "all" with "-all". So if they have +all it'll be -all. If they have ~all it'll be -all. Based on the SPF lib docs & my testing that's what appears to happen but I didn't really have a way to test it much without real connections coming in & it's just a test bed I was trying it on. I'm not sure if spf1 is force too with that setting.

One thing that needs to be confirmed is if it forces all spf to exactly "v=spf1 -all" (which wouldn't be too useful) or if it does what is assumed & that is just replace the "all" value. It'll take some testing of real incoming connections & viewing their spf records to see which is the case. (My old hacked up build I posted awhile back disabled +all by replacing it with -all. I contacted the author of the spf lib & he's the one who suggested to use the override value.)

Also the 2 spfdefault keys are for if they don't have spf defined where override is even if they do have it defined.
Bill
hMailServer build LIVE on my servers: 5.4-B2014050402
#hmailserver on FreeNode IRC https://webchat.freenode.net/?channels=#hmailserver
*** ABSENT FROM hMail! Those in IRC know how to find me if urgent. ***

marrco
New user
New user
Posts: 6
Joined: 2011-08-04 16:29

Re: LATEST EXPERIMENTAL BUILD - 5.4-B2014020801

Post by marrco » 2014-02-10 18:47

Bill48105 wrote:
marrco wrote:I had an other crash a few days ago, and now i updated to the new 2014/01/25 version.
Hey marco,
Were you having crashes before the new build? Did hmail crash or just log that error? If you can PM me more log lines around that error I'd appreciate it.
Thx
Bill
Hi Bill, sorry for the wrong info, hMailserver didn't crash, just stopped accepting new connections. Looking at the logs the error happened when i manually tried to restart it. That server is mainly used to do antivirus checks and spooling mail for a few internal MTA, and usually delivers a few thousand mail per day, and i realized about the problem after about one hour.

Code: Select all

"APPLICATION"	1684	"2014-02-04 11:24:20.360"	"SMTPDeliverer - Message 222323: Relaying to host my.internal.server.ip."
"APPLICATION"	1684	"2014-02-04 11:24:20.423"	"SMTPDeliverer - Message 222323: Message delivery thread completed."
"APPLICATION"	1684	"2014-02-04 11:24:44.634"	"SMTPDeliverer - Message 222324: Delivering message from prvs=3112eb0837=munged@senderdomain.it to mail@internaldomain.it. File: C:\Program Files (x86)\hMailServer\Data\{63D9008B-E8C7-419A-88DE-1FB42905A628}.eml"
"APPLICATION"	1688	"2014-02-04 11:24:47.863"	"SMTPDeliverer - Message 222325: Delivering message from prvs=611230c580=munged@senderdomain.it to user@otherdomaindomain.com. File: C:\Program Files (x86)\hMailServer\Data\{9FCF98D8-4B05-4814-8808-3689DDD61D90}.eml"
"APPLICATION"	1684	"2014-02-04 11:24:48.830"	"SMTPDeliverer - Message 222324: Relaying to host my.internal.server.ip."
"APPLICATION"	1684	"2014-02-04 11:24:48.893"	"SMTPDeliverer - Message 222324: Message delivery thread completed."
"APPLICATION"	1688	"2014-02-04 11:24:51.997"	"SMTPDeliverer - Message 222325: Relaying to host my.internal.server.ip."
"APPLICATION"	1688	"2014-02-04 11:24:52.122"	"SMTPDeliverer - Message 222325: Message delivery thread completed."
"APPLICATION"	1684	"2014-02-04 11:25:47.876"	"SMTPDeliverer - Message 221759: Delivering message from  to www-data@spammer-165-253.dtp.net.id. File: C:\Program Files (x86)\hMailServer\Data\{42A016B0-0B66-4021-8E40-3DB525B71099}.eml"
"APPLICATION"	1684	"2014-02-04 11:25:47.923"	"SMTPDeliverer - Message 221759: No mail servers could be found for the address www-data@spammer-165-253.dtp.net.id."
"APPLICATION"	1684	"2014-02-04 11:25:47.923"	"SMTPDeliverer - Message 221759: Message could not be delivered. Scheduling it for later delivery in 35 minutes."
"APPLICATION"	1684	"2014-02-04 11:25:47.923"	"SMTPDeliverer - Message 221759: Message delivery thread completed."
"APPLICATION"	1684	"2014-02-04 12:00:48.342"	"SMTPDeliverer - Message 221759: Delivering message from  to www-data@spammer-165-253.dtp.net.id. File: C:\Program Files (x86)\hMailServer\Data\{42A016B0-0B66-4021-8E40-3DB525B71099}.eml"
"APPLICATION"	1684	"2014-02-04 12:00:48.404"	"SMTPDeliverer - Message 221759: No mail servers could be found for the address www-data@spammer-165-253.dtp.net.id."
"APPLICATION"	1684	"2014-02-04 12:00:48.404"	"SMTPDeliverer - Message 221759: Message could not be delivered. Scheduling it for later delivery in 35 minutes."
"APPLICATION"	1684	"2014-02-04 12:00:48.404"	"SMTPDeliverer - Message 221759: Message delivery thread completed."
"APPLICATION"	1684	"2014-02-04 12:35:48.808"	"SMTPDeliverer - Message 221759: Delivering message from  to www-data@spammer-165-253.dtp.net.id. File: C:\Program Files (x86)\hMailServer\Data\{42A016B0-0B66-4021-8E40-3DB525B71099}.eml"
"APPLICATION"	1684	"2014-02-04 12:35:49.198"	"SMTPDeliverer - Message 221759: No mail servers could be found for the address www-data@spammer-165-253.dtp.net.id."
"APPLICATION"	1684	"2014-02-04 12:35:49.198"	"SMTPDeliverer - Message 221759: Message could not be delivered. Scheduling it for later delivery in 35 minutes."
"APPLICATION"	1684	"2014-02-04 12:35:49.198"	"SMTPDeliverer - Message 221759: Message delivery thread completed."
"APPLICATION"	1684	"2014-02-04 13:10:49.273"	"SMTPDeliverer - Message 221759: Delivering message from  to www-data@spammer-165-253.dtp.net.id. File: C:\Program Files (x86)\hMailServer\Data\{42A016B0-0B66-4021-8E40-3DB525B71099}.eml"
"APPLICATION"	1684	"2014-02-04 13:10:49.336"	"SMTPDeliverer - Message 221759: No mail servers could be found for the address www-data@spammer-165-253.dtp.net.id."
"APPLICATION"	1684	"2014-02-04 13:10:49.336"	"SMTPDeliverer - Message 221759: Message could not be delivered. Scheduling it for later delivery in 35 minutes."
"APPLICATION"	1684	"2014-02-04 13:10:49.336"	"SMTPDeliverer - Message 221759: Message delivery thread completed."
"APPLICATION"	1372	"2014-02-04 13:34:57.797"	"Stopping servers..."
"ERROR"	1824	"2014-02-04 13:35:00.449"	"Severity: 2 (High), Code: HM4208, Source: IOCPQueueWorkerTask::DoWork, Description: An unknown error occured while handling asynchronous requests. Error number: invalid vector<T> subscript"
"APPLICATION"	2288	"2014-02-04 13:36:38.870"	"Starting servers..."
"APPLICATION"	2288	"2014-02-04 13:36:39.790"	"Servers started."
"APPLICATION"	1768	"2014-02-04 13:36:53.300"	"SMTPDeliverer - Message 222326: Delivering message from munged@senderdomain to info@otherdomain.com. File: C:\Program Files (x86)\hMailServer\Data\{1634D008-9EFA-46D6-BDDD-372E906BC5E4}.eml"
"APPLICATION"	2944	"2014-02-04 13:36:53.736"	"SMTPDeliverer - Message 222327: Delivering message from campaign.270.3653@emails.ehmbars.co.uk to info@adomain.com. File: C:\Program Files (x86)\hMailServer\Data\{A309C4D2-3971-4D34-BA1F-89782C52A63C}.eml"
should you need more info just PM me

Bill48105
Developer
Developer
Posts: 6192
Joined: 2010-04-24 23:16
Location: Michigan, USA

Re: LATEST EXPERIMENTAL BUILD - 5.4-B2014020801

Post by Bill48105 » 2014-02-10 19:13

marrco wrote:Hi Bill, sorry for the wrong info, hMailserver didn't crash, just stopped accepting new connections. Looking at the logs the error happened when i manually tried to restart it. That server is mainly used to do antivirus checks and spooling mail for a few internal MTA, and usually delivers a few thousand mail per day, and i realized about the problem after about one hour.

** SNIP **

should you need more info just PM me
Yeah you can't be too surprised if you have errors during shutdown if hmail was busy but I'll look at that error to see if it gives clues as to why hmail stopped taking connections. So do you only have APPLICATION logging enabled or is the reason there is no SMTP/POP/IMAP/DEBUG/TCPIP logging because hmail wasn't accepting connections? I see the DNS lookup errors for that weird domain name but not sure if that is because your dns or internet is down or because of the weird domain.

Again if you have issues please create a new thread & tell you are using experimental build & version # & link back here if needed. We can't have long conversations on this release thread.
Thx
Bill
hMailServer build LIVE on my servers: 5.4-B2014050402
#hmailserver on FreeNode IRC https://webchat.freenode.net/?channels=#hmailserver
*** ABSENT FROM hMail! Those in IRC know how to find me if urgent. ***

User avatar
mattg
Moderator
Moderator
Posts: 22435
Joined: 2007-06-14 05:12
Location: 'The Outback' Australia

Re: LATEST EXPERIMENTAL BUILD - 5.4-B2014020801

Post by mattg » 2014-02-11 04:27

Hey Bill,

I run a few scripts :wink:

I have a rule that calls a custom function

Code: Select all

Function Check_eMail(oMessage)
	dim sRecipientName, sRecipientAddress
	sRecipientName = "Matt"
	sRecipientAddress = "matt@mydomain.com"

	oMessage.subject = "Checked by hMailserver script -" & oMessage.subject

	if left(oMessage.body,15) = "***************" then 
		oMessage.body = "********** Problem found **************" & vbcrlf & oMessage.body
		oMessage.addRecipient sRecipientName,sRecipientAddress
	else
		oMessage.body = "All checked OK" & VBCRLF & oMessage.body
		oMessage.subject = "ERROR -" & oMessage.subject
	end if
	oMessage.save
End Function
I can change a subject fine, but oMessage.body seems to be EMPTY.
The body is always "All checked OK" and nothing else.

If I comment out the if then else lines, and comment out the 'all checked ok' line, and just leave the 'problem found' line, the email body is ONLY "********** Problem found **************"

This is new behaviour since I have gone for the Alpha builds.

Matt
Just 'cause I link to a page and say little else doesn't mean I am not being nice.
https://www.hmailserver.com/documentation

Bill48105
Developer
Developer
Posts: 6192
Joined: 2010-04-24 23:16
Location: Michigan, USA

Re: LATEST EXPERIMENTAL BUILD - 5.4-B2014020801

Post by Bill48105 » 2014-02-11 07:09

mattg wrote:Hey Bill,

I run a few scripts :wink:

I have a rule that calls a custom function

Code: Select all

Function Check_eMail(oMessage)
	dim sRecipientName, sRecipientAddress
	sRecipientName = "Matt"
	sRecipientAddress = "matt@mydomain.com"

	oMessage.subject = "Checked by hMailserver script -" & oMessage.subject

	if left(oMessage.body,15) = "***************" then 
		oMessage.body = "********** Problem found **************" & vbcrlf & oMessage.body
		oMessage.addRecipient sRecipientName,sRecipientAddress
	else
		oMessage.body = "All checked OK" & VBCRLF & oMessage.body
		oMessage.subject = "ERROR -" & oMessage.subject
	end if
	oMessage.save
End Function
I can change a subject fine, but oMessage.body seems to be EMPTY.
The body is always "All checked OK" and nothing else.

If I comment out the if then else lines, and comment out the 'all checked ok' line, and just leave the 'problem found' line, the email body is ONLY "********** Problem found **************"

This is new behaviour since I have gone for the Alpha builds.

Matt
Hey Matt. No idea how that'd be the case can't say message body has been changed. you sure it worked ok with b1950? And it's blank for every email? Def odd.. I guess I can check the code where a rule calls the function & see if there's any code changes in that area but very unlikely. i'd sooner say it's something else & a coincidence you think it's the alpha builds but hey :D maybe we should chat about this in irc, email or pm's.
Thx
Bill
hMailServer build LIVE on my servers: 5.4-B2014050402
#hmailserver on FreeNode IRC https://webchat.freenode.net/?channels=#hmailserver
*** ABSENT FROM hMail! Those in IRC know how to find me if urgent. ***

User avatar
mattg
Moderator
Moderator
Posts: 22435
Joined: 2007-06-14 05:12
Location: 'The Outback' Australia

Re: LATEST EXPERIMENTAL BUILD - 5.4-B2014020801

Post by mattg » 2014-02-11 07:52

Definitely worked for > 2 years without change, and then stopped the day I loaded the Alpha.
I can send some messages to you if you like...

(This is for some automated scripts that send emails regularly, ie at least daily, detailing logs and other information)
Just 'cause I link to a page and say little else doesn't mean I am not being nice.
https://www.hmailserver.com/documentation

Bill48105
Developer
Developer
Posts: 6192
Joined: 2010-04-24 23:16
Location: Michigan, USA

Re: LATEST EXPERIMENTAL BUILD - 5.4-B2014020801

Post by Bill48105 » 2014-02-11 08:07

mattg wrote:Definitely worked for > 2 years without change, and then stopped the day I loaded the Alpha.
I can send some messages to you if you like...

(This is for some automated scripts that send emails regularly, ie at least daily, detailing logs and other information)
Well you hadn't had B1950 for 2 years ;) Is why I wondered if you were sure B1950 worked OK. Or if you restored you B1950 bin & see if it works again. i can't think of any reason for it to stop working between 1950 & Jan alpha. btw did you drop ALL files from the 7z into bin or just hmailserver.exe? when there are COM changes all files need to be replaced. (Well all in the 7z. not the ini for example)
hMailServer build LIVE on my servers: 5.4-B2014050402
#hmailserver on FreeNode IRC https://webchat.freenode.net/?channels=#hmailserver
*** ABSENT FROM hMail! Those in IRC know how to find me if urgent. ***

User avatar
mattg
Moderator
Moderator
Posts: 22435
Joined: 2007-06-14 05:12
Location: 'The Outback' Australia

Re: LATEST EXPERIMENTAL BUILD - 5.4-B2014020801

Post by mattg » 2014-02-11 14:48

Bill48105 wrote:Is why I wondered if you were sure B1950 worked OK.
Yep
Bill48105 wrote:Or if you restored you B1950 bin & see if it works again.
yep
Bill48105 wrote:i can't think of any reason for it to stop working between 1950 & Jan alpha. btw did you drop ALL files from the 7z into bin or just hmailserver.exe? when there are COM changes all files need to be replaced. (Well all in the 7z. not the ini for example)
Yep, and then updated my ini...

Just checked, reverting back to my old INI still has the same issue with the alpha build.
Just 'cause I link to a page and say little else doesn't mean I am not being nice.
https://www.hmailserver.com/documentation

User avatar
mattg
Moderator
Moderator
Posts: 22435
Joined: 2007-06-14 05:12
Location: 'The Outback' Australia

Re: LATEST EXPERIMENTAL BUILD - 5.4-B2014020801

Post by mattg » 2014-02-14 05:17

encoding issue

http://www.hmailserver.com/forum/viewto ... 58#p158258

All sorted
Thanks to Bill for troubleshooting via IRC & to Kriztan for the idea in another thread
Just 'cause I link to a page and say little else doesn't mean I am not being nice.
https://www.hmailserver.com/documentation

User avatar
mattg
Moderator
Moderator
Posts: 22435
Joined: 2007-06-14 05:12
Location: 'The Outback' Australia

Re: LATEST EXPERIMENTAL BUILD - 5.4-B2014020801

Post by mattg » 2014-02-14 05:46

Not quite all sorted.

Some of these mail that coming to me, aren't generated by my code initially.
One example is the automated emails that are sent from ShadowProtect.
Just 'cause I link to a page and say little else doesn't mean I am not being nice.
https://www.hmailserver.com/documentation

Bill48105
Developer
Developer
Posts: 6192
Joined: 2010-04-24 23:16
Location: Michigan, USA

Re: LATEST EXPERIMENTAL BUILD - 5.4-B2014021401

Post by Bill48105 » 2014-02-14 08:31

NEW BUILD POSTED
2014-02-08 5.4-B2014020801
* IMPORTANT: This build has a LOT of extra debug logging but NOT shown by default. [Settings]LogLevel=10 for some extra to 100 for extremely verbose
* FIX: Blank oMessage.body if MIME but no charset after UTF8 indexing patch
* ADDED: Extra debug logging for adding attachments in scripts LogLevel >9
hMailServer build LIVE on my servers: 5.4-B2014050402
#hmailserver on FreeNode IRC https://webchat.freenode.net/?channels=#hmailserver
*** ABSENT FROM hMail! Those in IRC know how to find me if urgent. ***

prisma
Senior user
Senior user
Posts: 325
Joined: 2010-07-09 13:16

Re: LATEST EXPERIMENTAL BUILD - 5.4-B2014021401

Post by prisma » 2014-02-14 16:49

5.4-B2014021401 runs in production environment. Nothing to report until now, everything works. But:

I had debug logging (accidentally) switched on (normal debug level) and noticed following:

Code: Select all

"DEBUG"	3396	"2014-02-14 11:09:28.255"	"TCPConnection::Write non ssl"
"DEBUG"	1556	"2014-02-14 11:09:28.255"	"TCPConnection::Write non ssl"
and variations of read/write and non/ssl. Such lines are repeating 10thousands time over a few seconds and within milliseconds and in big repeating blocks.

Is this intended and/or possibly normal logging behaviour?
Or an error/overreaction within a log function?
Or is it a sign for a deeper problem we should follow up?

User avatar
mattg
Moderator
Moderator
Posts: 22435
Joined: 2007-06-14 05:12
Location: 'The Outback' Australia

Re: LATEST EXPERIMENTAL BUILD - 5.4-B2014021401

Post by mattg » 2014-02-14 16:54

I had to turn logging down to level 1 or 2 to get normal logs on that version.
Bill has fixed that with the latest build
Just 'cause I link to a page and say little else doesn't mean I am not being nice.
https://www.hmailserver.com/documentation

Bill48105
Developer
Developer
Posts: 6192
Joined: 2010-04-24 23:16
Location: Michigan, USA

Re: LATEST EXPERIMENTAL BUILD - 5.4-B2014021401

Post by Bill48105 » 2014-02-14 17:24

prisma wrote:5.4-B2014021401 runs in production environment. Nothing to report until now, everything works. But:

I had debug logging (accidentally) switched on (normal debug level) and noticed following:

Code: Select all

"DEBUG"	3396	"2014-02-14 11:09:28.255"	"TCPConnection::Write non ssl"
"DEBUG"	1556	"2014-02-14 11:09:28.255"	"TCPConnection::Write non ssl"
and variations of read/write and non/ssl. Such lines are repeating 10thousands time over a few seconds and within milliseconds and in big repeating blocks.

Is this intended and/or possibly normal logging behaviour?
Or an error/overreaction within a log function?
Or is it a sign for a deeper problem we should follow up?
mattg wrote:I had to turn logging down to level 1 or 2 to get normal logs on that version.
Bill has fixed that with the latest build
Yeah as mattg says I changed how the loglevel works in the last build (Feb 14) vs the previous few builds. LogLevel defaults to 9 if not defined. In the Jan & earlier Feb builds 9 showed a LOT of debug logging but not all the extra logging. That meant you had to set <7 (or as matt says even as low as 2 or 1) to hide some of the new logging. Now with Feb 14 build 9 shows NO EXTRA logging by default. If you want SOME extra logging set it to 10 or 99 to show crazy verbose logging on Feb 14 build. Or on Jan & earlier Feb builds set it to 2 or 1 to hide all the extra logging.
Bill
hMailServer build LIVE on my servers: 5.4-B2014050402
#hmailserver on FreeNode IRC https://webchat.freenode.net/?channels=#hmailserver
*** ABSENT FROM hMail! Those in IRC know how to find me if urgent. ***

prisma
Senior user
Senior user
Posts: 325
Joined: 2010-07-09 13:16

Re: LATEST EXPERIMENTAL BUILD - 5.4-B2014021401

Post by prisma » 2014-02-14 18:27

Understand. This means I can switch on debug log again (if I want or need) without setting a lower level, because it defaults now to 9 and this means no extra debug logging. Except I need extra logging :) Sorry, on a Friday evening I'm a little bit slower... Have a nice weekend!

Bill48105
Developer
Developer
Posts: 6192
Joined: 2010-04-24 23:16
Location: Michigan, USA

Re: LATEST EXPERIMENTAL BUILD - 5.4-B2014021401

Post by Bill48105 » 2014-02-14 18:49

prisma wrote:Understand. This means I can switch on debug log again (if I want or need) without setting a lower level, because it defaults now to 9 and this means no extra debug logging. Except I need extra logging :) Sorry, on a Friday evening I'm a little bit slower... Have a nice weekend!
Yupper. Sorry about that I had idea of how i was doing loglevel but needed to balance existing debug logging with the new stuff for test builds. Hopefully the new levels work out ok for people or I'll tweak again.
Bill
hMailServer build LIVE on my servers: 5.4-B2014050402
#hmailserver on FreeNode IRC https://webchat.freenode.net/?channels=#hmailserver
*** ABSENT FROM hMail! Those in IRC know how to find me if urgent. ***

User avatar
mattg
Moderator
Moderator
Posts: 22435
Joined: 2007-06-14 05:12
Location: 'The Outback' Australia

Re: LATEST EXPERIMENTAL BUILD - 5.4-B2014021401

Post by mattg » 2014-03-04 07:47

Bill,

Just gotta say that I've been on B2014021401 since that day, and rock solid mate.

Not missing a beat.
I have StartTLS set for a custom incoming SMTP port, and it is working fine...
My scripts are all working as expected and my error logs have shrink immensely from b1950

Matt
Just 'cause I link to a page and say little else doesn't mean I am not being nice.
https://www.hmailserver.com/documentation

Bill48105
Developer
Developer
Posts: 6192
Joined: 2010-04-24 23:16
Location: Michigan, USA

Re: LATEST EXPERIMENTAL BUILD - 5.4-B2014021401

Post by Bill48105 » 2014-03-04 08:44

mattg wrote:Bill,

Just gotta say that I've been on B2014021401 since that day, and rock solid mate.

Not missing a beat.
I have StartTLS set for a custom incoming SMTP port, and it is working fine...
My scripts are all working as expected and my error logs have shrink immensely from b1950

Matt
Great to hear man! :) Thx for the feedback
Bill
hMailServer build LIVE on my servers: 5.4-B2014050402
#hmailserver on FreeNode IRC https://webchat.freenode.net/?channels=#hmailserver
*** ABSENT FROM hMail! Those in IRC know how to find me if urgent. ***

Bill48105
Developer
Developer
Posts: 6192
Joined: 2010-04-24 23:16
Location: Michigan, USA

Re: LATEST EXPERIMENTAL BUILD - 5.4-B2014040801

Post by Bill48105 » 2014-04-08 19:17

2014-04-08 5.4-B2014040801
* IMPORTANT: This build has a LOT of extra debug logging but NOT shown by default. [Settings]LogLevel=10 for some extra to 100 for extremely verbose
* URGENT: Heartbleed zero day openSSL fix http://www.zdnet.com/heartbleed-serious ... 000028166/
* Upated hmailserver to openssl-1.0.1g
* ONLY MINIMALLY TESTED. VULNERABILITY ERADICATION UNTESTED. Just built openssl-1.01.1g & incorporated into this build.
hMailServer build LIVE on my servers: 5.4-B2014050402
#hmailserver on FreeNode IRC https://webchat.freenode.net/?channels=#hmailserver
*** ABSENT FROM hMail! Those in IRC know how to find me if urgent. ***

Bill48105
Developer
Developer
Posts: 6192
Joined: 2010-04-24 23:16
Location: Michigan, USA

Re: LATEST EXPERIMENTAL BUILD - 5.4-B2014040801

Post by Bill48105 » 2014-04-08 19:30

Also posted non-static linked build so openssl dll's can be used if needed. Obviously weird things can happen if you use incompatible DLL's so know what you are doing & only use if needed.
http://www.hmailserver.com/forum/viewto ... 10&t=21420
hMailServer build LIVE on my servers: 5.4-B2014050402
#hmailserver on FreeNode IRC https://webchat.freenode.net/?channels=#hmailserver
*** ABSENT FROM hMail! Those in IRC know how to find me if urgent. ***

percepts
Senior user
Senior user
Posts: 5282
Joined: 2009-10-20 16:33
Location: Sceptred Isle

Re: LATEST EXPERIMENTAL BUILD - 5.4-B2014040801

Post by percepts » 2014-04-08 21:55

Bill,

Now martin has released B1951 production release where does that leave us with regard to experimental releases?
Will there need to be new experimental release based on B1951 or can put already released eperimental versions on top of B1951? Does 1951 include any/all of the experimental features?

User avatar
mattg
Moderator
Moderator
Posts: 22435
Joined: 2007-06-14 05:12
Location: 'The Outback' Australia

Re: LATEST EXPERIMENTAL BUILD - 5.4-B2014040801

Post by mattg » 2014-04-08 22:59

percepts wrote: Does 1951 include any/all of the experimental features?
Or other fixes that you have developed...Like the UTF-8 Indexing, the IOCP fixes etc
Just 'cause I link to a page and say little else doesn't mean I am not being nice.
https://www.hmailserver.com/documentation

Post Reply