Re: [vchkpw] Any 5.4.28 or 5.5 ETA?
-BEGIN PGP SIGNED MESSAGE- Hash: SHA1 Wouter van der Schagt wrote: Hi all, I know it has been a while, but a few months ago, I commented on an error message when deleting a domain using vdeldomain in the 5.5 branch. The error message was: Warning: Failed while attempting to delete domain from auth backend. I believe at that time, the cause of the message has been identified by mr. Brookings. Just to let you know. On a clean install of vpopmail 5.4.25. This message also appears. However only it seems because it hadn't been used yet (the server). I'm not sure I remember this issue, but I don't think there's anything in the tracker about it. That's *probably* because I resolved it. Is there an ETA on the release of the 5.4.28 or 5.5 branch? Most notably the 5.5 branch. We are anxious to further experiment with it in more production environments. Is there anything we can help with? I'll be releasing a 5.4.28 release candidate here shortly. I was out of the office on a short vacation, but I'm back now, and will be returning to work on vpopmail and qmailadmin. The 5.5 branch's time frame depends greatly on 5.4.28's time line. When 5.4.28 goes stable, the final version of the 5.4 tree, 5.5 will become the development version looking for a stable release. - -- /* Matt Brookings m...@inter7.com GnuPG Key FAE0672C Software developer Systems technician Inter7 Internet Technologies, Inc. (815)776-9465 */ -BEGIN PGP SIGNATURE- Version: GnuPG v1.4.9 (GNU/Linux) Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org iEYEARECAAYFAkpvERoACgkQIwet2/rgZyxmcgCdECEIFGAuHiK/LIagyhD8DXsk hYwAnjfpxTM/kYOrGjBT5xhz+cEDgS2Y =TSPx -END PGP SIGNATURE-
Re: [vchkpw] 5.4.28 test remarks
-BEGIN PGP SIGNED MESSAGE- Hash: SHA1 Wouter van der Schagt wrote: When using vpopmail 5.4.28-devel. I'm getting the following error message: configure: error: Unable to find your MySQL inc dir, specify --enable-incdir. This message appeared when the development files for MySQL were not installed, but the directory existed. Installing the necessary files worked, but perhaps the message could be a little different because the option 'was' specified. Anyway, this is only a minor issue. Because the configure script uses existing autoconf tool calls for determining if the development files for MySQL are existing, it wouldn't make sense to modify this. The test is testing if it can link to MySQL at standard locations. If you don't have the development files, it matters little whether the directory exists or not. However, I do agree it could say something along the lines of 'Unable to find MySQL development headers', rather than that it could not find the directory. --- 2 --- The status at the end of the configure, shows 5.4.27, even though it is 5.4.28, again - minor. Really? Ha! I'll fix that :) --- 3 --- In 5.5 I noticed that the tcp.smtp file was moved to /home/vpopmail/etc/tcp.smtp instead of /etc/tcp.smtp. Is this the new default behavior in 5.4.28 as well? This would probably mean that people will have to update their existing /var/qmail/bin/qmailctl file for the cdb command to reload the rules? The 5.5 branch is working in FHS compliance, and it's a bit complex because so much of vpopmail has been built on /home/vpopmail. Normally vpopmail looks for an existing tcp.smtp file and then favors that over it's desired /home/vpopmail/etc location. Adding the FHS favoring in also caused some pathing issues with the tcp.smtp file. Just force it for now. This will be worked out as 5.5 becomes the devel version. The previously reported error message: Warning: Failed while attempting to delete domain from auth backend also appears in 5.4.28 when adding and immediately deleting a domain. I believe the cause is also the same as in the 5.4.25 and 5.5 branch. If I remember correctly a strace revealed a failing SQL query on a non existing table. I can look it up if necessary, just let me know. Please do because I can't remember what happened with this. What is the default behavior of the newly implemented domain quota's in 5.4.28? is it on or off by default, and if on, how can it be controlled? Domain quotas are *not* on by default in 5.4.28 because they haven't been in so long. Also, 5.5 differs greatly in quota handling, and 5.4.28 does not have these changes, so I felt it best to leave 5.4.28 with domain quotas off by default. Last question: Is there any major change that could break an existing 5.4.25 installation? I'm considering to upgrade a production server to 5.4.28 status in order to test the usage daemon, the 2 GB per mailbox fix and possibly the domain quotas. I'm not aware of any at this time. The point of 5.4.28 was to fix up some big problems with the quota system, solve any remaining large bugs, and finally, to release a final 5.4 tree version so the new 5.5 tree could go development with it's *many*, *big* changes. - -- /* Matt Brookings m...@inter7.com GnuPG Key FAE0672C Software developer Systems technician Inter7 Internet Technologies, Inc. (815)776-9465 */ -BEGIN PGP SIGNATURE- Version: GnuPG v1.4.9 (GNU/Linux) Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org iEYEARECAAYFAkpvEvkACgkQIwet2/rgZyyJBwCePvmPKroW7dZwN1oxRLIMIETd 6PIAoIuZvNavdlsmN2MIImm4V43YV18E =gMl0 -END PGP SIGNATURE-
Re: [vchkpw] 5.4.28 - vusaged - error - deferral: Aack,_child_crashed._(#4.3.0)
-BEGIN PGP SIGNED MESSAGE- Hash: SHA1 Wouter van der Schagt wrote: Good morning all, I have installed Vpopmail 5.4.28 on a few production mail servers and noticed a problem. As long as vusaged is running, the following error message appears in /var/log/qmail/current: deferral: Aack,_child_crashed._(#4.3.0) which causes all local email to stay in the queue 'forever'. Shutting down the vusage daemon has the desired effect that email is being delivered. The message then becomes: @40004a6d6b3c347e31bc delivery 951: success: client_connect:_connect_failed:_111/client_connect:_connect_failed:_111/client_connect:_connect_failed:_111/client_connect:_connect_failed:_111/did_0+0+1/ This message appears because it cannot connect to the vusage daemon of course. But the email is deliverd. Any idea why I get child crashed when trying to deliver local emails? Unfortunately I have not seen this on any of the production servers I have 5.4.28 on. It definitely sounds like an issue on the vusage client API end, but from here, I can't tell. I would try using vdelivermail by hand, or the vusagec binary by hand, and see if that crashes. - -- /* Matt Brookings m...@inter7.com GnuPG Key FAE0672C Software developer Systems technician Inter7 Internet Technologies, Inc. (815)776-9465 */ -BEGIN PGP SIGNATURE- Version: GnuPG v1.4.9 (GNU/Linux) Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org iEYEARECAAYFAkpvE3QACgkQIwet2/rgZyxKIgCcDRRRntWAuGnTRgQ8ynB/QBfP 5pwAoIaN9DW53gt2R82e6g3nQaMpNC+O =R9HW -END PGP SIGNATURE-
Re: [vchkpw] 5.4.28 test remarks
Good morning, The 5.5 branch is working in FHS compliance, and it's a bit complex because so much of vpopmail has been built on /home/vpopmail. Normally vpopmail looks for an existing tcp.smtp file and then favors that over it's desired /home/vpopmail/etc location. Adding the FHS favoring in also caused some pathing issues with the tcp.smtp file. Just force it for now. This will be worked out as 5.5 becomes the devel version. Allright, actually I remember you saying this before. I forgot. I used: --enable-tcpserver-file=/etc/tcp.smtp for now. Warning: Failed while attempting to delete domain from auth backend Please do because I can't remember what happened with this. Allright, below is a snippet of the email from before: (it seems to happen after a clean install, at which time, no logins have been recorded yet, perhaps the table is not yet created? --- I performed a strace on the command vdeldomain and found: write(4, 1\0\0\0\3delete from lastauth where d..., 53) = 53 read(4, 0\0\0\1\377z\4#42S02Table 'vpopmail.last..., 16384) = 52 write(2, Warning: Failed while attempting ..., 68Warning: Failed while attempting to delete domain from auth backend ) = 68 Manual investigation showed that there is no table lastauth in the vpopmail database. Should this have been created on install ? or is this created when somebody logs in? The error doesn't appear to be fatal, but it is ugly. --- Sincerely, - Wouter van der Schagt !DSPAM:4a6fe34332714395816559!
Re: [vchkpw] 5.4.28 - vusaged - error - deferral: Aack,_child_crashed._(#4.3.0)
Good morning, I would try using vdelivermail by hand, or the vusagec binary by hand, and see if that crashes. Using vusagec by hand returns the total number of messages and number of bytes used. Using vdelivermail by hand, results in: vdelivermail: no EXT environment varilable, the command used is: /home/vpopmail/bin/vdelivermail wou...@vdschagt.com /tmp/testmsg.eml Now I realize that command may be wrong, however have no idea how to use it properly? How can I test it manually? Sincerely, - Wouter van der Schagt !DSPAM:4a6fe3f732711631481000!
[vchkpw] Vusaged domain total number of messages always 0
Good morning again On another production server, while I still get the the child crashed problem, I also noticed the following: --- ms3:~# /home/vpopmail/bin/vusagec wou...@vdschagt.com wou...@vdschagt.com: 1965201 byte(s) in 246 message(s) --- As expected. However when looking at the total of a domain: ms3:~# /home/vpopmail/bin/vusagec @vdschagt.com vdschagt.com: 2210342 byte(s) in 0 message(s) --- Notice the 0 messages in the total of a domain quota. Is this expected behavior? Sincerely, - Wouter van der Schagt !DSPAM:4a6fe4af32714294588090! Emoticon1.gif