HELIX 7.1.1 ported to FreeBSD

Printable View

Show 40 post(s) from this thread on one page

11-21-2011, 06:12 AM
ralenekov

The upgrade from 6.0.13 went just fine :D and the mail server is working perfectly stable for more than a month now.

The only modification that I did was to update clamav to 0.97.3 after the first week of usage.

I have another problem which seems to be known in the 7.x versions of Zimbra: the server has high CPU usage by the java processes. It is stated that this bug is fixed in the patchset for the 7.1.3 version but some people are reporting the same problem even after applying the patchset to 7.1.3.

So I am waiting for an update patch... and the new port :rolleyes:

On 6.x there was one java process, and on 7.x there are two, which are terribly slow on "zmcontrol stop". Is this the same problem I just mentioned or there is a known setting for this?
02-27-2012, 11:10 AM
solko

Helix-714

HELIX-714

please, see the patch notes. this is a major change.

FreeBSD-8 and FreeBSD-9 packages.
06-03-2012, 01:05 PM
ralenekov

Great build as always. Thank you very much! :)

The upgrade from 7.1.1 to 7.1.4 went smoothly.

[zimbra@mail ~]$ zmcontrol -v
Release 7.1.4_GA_2565.FreeBSD_amd64 FreeBSD_amd64 FOSS edition.
[zimbra@mail ~]$

But I came across two strange "bugs":

The first one is that the function "View mail" in the admin GUI is kind of broken. It always redirects to port 80 (https://mail.server.com:80). It appends :80 to the address every time. If you delete the ":80" stuff in the browser - the user mail opens fine, so I think it's not a problem with the auth tokens or something.

I set this (followed by several zmcontrol restarts) but it makes no difference:
[zimbra@mail ~]$ zmprov gd 'server.com' | grep Public
zimbraPublicServiceHostname: mail.server.com
zimbraPublicServicePort: 443
zimbraPublicServiceProtocol: https
[zimbra@mail ~]$

zmtlsctl mixed/redirect don't seem to make a difference either...
It seems like it is hardcoded somewhere...

The second on is that the web server statistics show that every service is down (red cross). Everywhere in the system is reported that all zimbra services are up, including in the log files. I tried /opt/zimbra/libexec/zmsyslogsetup and my syslogd.conf is ok but it didn't help... My cron entry is scheduled for every 10 min. Do I have to use rsyslogd in order for these stats to work?
06-03-2012, 05:39 PM
solko

Quote:

Originally Posted by ralenekov

Great build as always. Thank you very much! :)

The upgrade from 7.1.1 to 7.1.4 went smoothly.

[zimbra@mail ~]$ zmcontrol -v
Release 7.1.4_GA_2565.FreeBSD_amd64 FreeBSD_amd64 FOSS edition.
[zimbra@mail ~]$

But I came across two strange "bugs":

The first one is that the function "View mail" in the admin GUI is kind of broken. It always redirects to port 80 (https://mail.server.com:80). It appends :80 to the address every time. If you delete the ":80" stuff in the browser - the user mail opens fine, so I think it's not a problem with the auth tokens or something.

I set this (followed by several zmcontrol restarts) but it makes no difference:
[zimbra@mail ~]$ zmprov gd 'server.com' | grep Public
zimbraPublicServiceHostname: mail.server.com
zimbraPublicServicePort: 443
zimbraPublicServiceProtocol: https
[zimbra@mail ~]$

zmtlsctl mixed/redirect don't seem to make a difference either...
It seems like it is hardcoded somewhere...

The second on is that the web server statistics show that every service is down (red cross). Everywhere in the system is reported that all zimbra services are up, including in the log files. I tried /opt/zimbra/libexec/zmsyslogsetup and my syslogd.conf is ok but it didn't help... My cron entry is scheduled for every 10 min. Do I have to use rsyslogd in order for these stats to work?

sadly, no idea about the first but might be related to the upgrade 7.1.1 -> 7.1.4. i don't think that fresh installation was doing this but i will check.

as for the second, it might be (probably is) related to swatch. could you please delete swatch package from base system? might not do a thing but worth a try.
also, please verify that all services are running via 'zmcontrol status', i think there was some startup issue with zmlogstats.
06-04-2012, 03:03 AM
ralenekov

Some info considering the zmlogger problem:

I checked with or without the swatch package - no difference.

Code:

[zimbra@mail ~/log]$ tail /var/log/zimbra-stats.log Jun 4 12:40:07 mail zimbramon[72908]: 72908:info: 2012-06-04 12:40:01, STATUS: mail.server.com: antispam: Running Jun 4 12:40:07 mail zimbramon[72908]: 72908:info: 2012-06-04 12:40:01, STATUS: mail.server.com: antivirus: Running Jun 4 12:40:07 mail zimbramon[72908]: 72908:info: 2012-06-04 12:40:01, STATUS: mail.server.com: ldap: Running Jun 4 12:40:07 mail zimbramon[72908]: 72908:info: 2012-06-04 12:40:01, STATUS: mail.server.com: logger: Running Jun 4 12:40:07 mail zimbramon[72908]: 72908:info: 2012-06-04 12:40:01, STATUS: mail.server.com: mailbox: Running Jun 4 12:40:07 mail zimbramon[72908]: 72908:info: 2012-06-04 12:40:01, STATUS: mail.server.com: mta: Running Jun 4 12:40:07 mail zimbramon[72908]: 72908:info: 2012-06-04 12:40:01, STATUS: mail.server.com: snmp: Running Jun 4 12:40:07 mail zimbramon[72908]: 72908:info: 2012-06-04 12:40:01, STATUS: mail.server.com: spell: Running Jun 4 12:40:07 mail zimbramon[72908]: 72908:info: 2012-06-04 12:40:01, STATUS: mail.server.com: stats: Running Jun 4 12:40:07 mail zimbramon[72908]: 72908:info: 2012-06-04 12:40:01, STATUS: mail.server.com: zmconfigd: Running [zimbra@mail ~/log]$

Code:

[root@mail /opt/zimbra]# cat /opt/zimbra/log/zmlogswatch.out zmrrdfetch server pid: 72858 bind: Address already in use at /opt/zimbra/libexec/zmlogger line 471. Schema version not found, upgrading to version 4 at /opt/zimbra/libexec/zmlogger line 412. Use of uninitialized value $hostid in exists at /opt/zimbra/libexec/zmlogger line 128, <> line 1. Use of uninitialized value $hostid in hash element at /opt/zimbra/libexec/zmlogger line 128, <> line 1. Use of uninitialized value $hostid in hash element at /opt/zimbra/libexec/zmlogger line 129, <> line 1. Use of uninitialized value $hostid in hash element at /opt/zimbra/libexec/zmlogger line 141, <> line 1. Use of uninitialized value $hostid in hash element at /opt/zimbra/libexec/zmlogger line 142, <> line 1. Use of uninitialized value $hostid in exists at /opt/zimbra/libexec/zmlogger line 240, <> line 1. Use of uninitialized value $hostid in hash element at /opt/zimbra/libexec/zmlogger line 240, <> line 1. 19 at /opt/zimbra/libexec/zmlogger line 265, <> line 1. Issuing rollback() due to DESTROY without explicit disconnect() of DBD::SQLite::db handle dbname=/opt/zimbra/logger/db/data/logger.sqlitedb at /opt/zimbra/libexec/zmlogger line 265, <> line 1. [root@mail /opt/zimbra]#

Code:

[zimbra@mail ~]$ zmcontrol status Host mail.server.com antispam Running antivirus Running ldap Running logger Stopped zmlogswatchctl is not running mailbox Running mta Running snmp Running spell Running stats Running zmconfigd Running [zimbra@mail ~]$

Code:

[zimbra@mail ~]$ /opt/zimbra/libexec/zmrrdfetch -f zmmtastats Can't call method "bind_param" on an undefined value at /opt/zimbra/libexec/zmrrdfetch line 62. [zimbra@mail ~]$ /opt/zimbra/libexec/zmrrdfetch -c -f zmmtastats Can't call method "bind_param" on an undefined value at /opt/zimbra/libexec/zmrrdfetch line 62.

I googled some solutions but nothing helped... :(
06-04-2012, 03:52 AM
solko

Quote:

Originally Posted by ralenekov

Some info considering the zmlogger problem:
I googled some solutions but nothing helped... :(

in that case i fear that it is a direct result of changes i had to incorporate and that the upgrade procedure is somehow unable to cope with it.
perhaps, deleting all the application stuff while leaving data/config intact should do the trick but i am in no position to suggest anything :/. i always create a snapshot before upgrade and do a rollback if anything goes wrong. which is probably not an option for you.

basically, in previous versions some parts of logger were reused from base system. in later versions logger (and almost all other zimbra parts) is no longer dependent on base and all its parts are 'zimbra-as-is' (with some tweaks to run on bsd).

BUT. i think that you should completely delete the stats database, please google it out and/or try:
http://www.zimbra.com/forums/adminis...-6-0-1-ne.html.
i vaguely recall i had done that in some point of my life :-).

please, let me know on skype. i am kinda busy atm and this forum approach is slow and unpractical ;).
06-06-2012, 07:40 AM
ralenekov

Quote:

Originally Posted by solko

BUT. i think that you should completely delete the stats database, please google it out and/or try:
http://www.zimbra.com/forums/adminis...-6-0-1-ne.html.
i vaguely recall i had done that in some point of my life :-).

I did that before my previos post. It didn't make a difference.

It seems that this is some kind of bug:
http://www.zimbra.com/forums/adminis...l-problem.html

Can I just take the latest patch for 7.1.4:
http://files2.zimbra.com/downloads/7....4_GA_2568.tgz
... and overwrite the jar files? Will it work out or should I wait for you to release 7.1.4 2568 build or newer?

Thanks.
06-11-2012, 02:42 AM
ralenekov

The solution to this problem:

Quote:

Originally Posted by ralenekov

The first one is that the function "View mail" in the admin GUI is kind of broken. It always redirects to port 80 (https://mail.server.com:80). It appends :80 to the address every time.

1: Updated zimbra to 7.1.4 PL1 (2568)

Quote:

[zimbra@mail /root]$ zmcontrol -v
Release 7.1.4_GA_2565.FreeBSD_amd64 FreeBSD_amd64 FOSS edition, Patch 7.1.4_P1.
[zimbra@mail /root]$

It didn't help.

2: I don't know if it is a bug or it just works that way.

I am using "zmtlsctl redirect" but when I checked:

Quote:

[zimbra@mail /root]$ zmprov gs `zmhostname` | grep zimbraMailPort
zimbraMailPort: 80
[zimbra@mail /root]$

Quote:

[zimbra@mail /root]$ zmprov ms mail.chaosgroup.com ZimbraMailPort 443
[zimbra@mail /root]$ zmprov gs `zmhostname` | grep zimbraMailPort
[zimbra@mail /root]$ zimbraMailPort: 443

... and the "view mail" function works as expected now.

I will report when I come to solution for second "bug".
06-29-2012, 01:03 AM
solko

Helix-720

HELIX-720 on FreeBSD - Zimbra :: Wiki

this release follows the same logic as the previous one (therefore HELIX-714 build/install guide applies). if you want to build yourself just fetch the prepared source and builddeps.

i linked.. yet again.. against local libpng. please install png via:
pkg_add -r png
else logger won't run properly.

i linked.. yet again.. against local freetype2. please install freetype2 via:
pkg_add -r freetype2
else logger won't run properly.

...
07-24-2012, 12:43 AM
Derzelis

7.2.0_GA_2678 patch

Hi Solko ,
We need to patch and update regurarly the Zimbra releases for FreeBsd.
Up to now , the method described here Zimbra on FreeBSD - Zimbra :: Wiki worked up to 714 with minor tweakings (jail settings).
We are confused by 720 approach. Our uole is understanding is that the patch has been merged into the sourcecode , and the whole is available for downlaoding.
This is great for people willing to install , but not good for people willing to compile the code .
Would it be possible to put the patch for 720 just as it was the case for previous versions (separate from zimbra source code) ?
Like this , more people could contribute if needed.

Some remarks concerning 7.2.0_GA_2678 :
- on freebsd9-current , nginx fail to compile
- if we install the precompiled version several perl packages are still needed postinstall (mainly for amavis ) p5-Convert-UUlib , p5-HTML-Parser , p5-NetAddr-IP , p5-Crypt-OpenSSL-RSA , p5-IO-Compress , p5-Unix-Syslog

Quote:

Originally Posted by solko

HELIX-720 on FreeBSD - Zimbra :: Wiki

this release follows the same logic as the previous one (therefore HELIX-714 build/install guide applies). if you want to build yourself just fetch the prepared source and builddeps.

i linked.. yet again.. against local libpng. please install png via:
pkg_add -r png
else logger won't run properly.

i linked.. yet again.. against local freetype2. please install freetype2 via:ura
pkg_add -r freetype2
else logger won't run properly.

...

Show 40 post(s) from this thread on one page