Show Posts

This section allows you to view all posts made by this member. Note that you can only see posts made in areas you currently have access to.


Messages - uacnix

Pages: [1]
1
Email and Groupware / Re: Postgres process hogging HDD 90% of the time.
« on: January 11, 2017, 02:49:57 am »
We have about 150-200 users, mostly older people "I click deze icons and dose mails show up" computer illiterates.
EDIT:
I did the following in terminal:
Code: [Select]
sudo su postgres
psql -c "select * from pg_stat_activity"
see that mess:

Think - "Hell, why not? I'm not in the office tomorrow, so they'll have to live with it if anything goes wrong"
Code: [Select]
psql -c "select pg_terminate_backend(4536)"Killed some douchebag query which was hogging the HDD all the time, server load levels went right away back to normal (e.g. to 0.08 from 2.60)
Saved the world, day and my ass since I don't have to move it to a new machine, just backup/clone the drive in Acronis just in case and do zentyal upgrade right away from the webinterface.

Next thing I gotta solve is:
Quote
How do i protect this server from spoofing our own emails, like my email is it@domain.com, and i get mails from it@domain.com, also how do i add more sophisticated rules into spamassassin's local.cf and make that actually work, since "module reload just grabs the list of whitelist/blacklist emails and does plain whitelist_from/blacklist_from, where I can use whitelist_from_rcvd or shortcircuiting for lower CPU usage"
But thats another story :3 Thank You all for your help guys!

2
Email and Groupware / Re: Postgres process hogging HDD 90% of the time.
« on: January 09, 2017, 11:00:29 pm »
Hello,

Sorry I didn't reply, I was on holiday and kinda forgot about the work.
Yesterday I set up Zentyal 4.2 on some free PC I had laying around (Intel C2d E7200 -old server has Xeon X3430, yet with better drive - WD Black 1TB, and 2 gigs of ram instead of pathetic 1gb).
On the old server, the accounts created in the webinterface are also created in /home/<accountname>, and the actual mails are put into /var/vmail/domainname/<accountname>, where they have cur, new, temp and some .Drafts and other folders - almost every mailfolder seems to have those subfolders, also every folder has dovecot-uidlist and dovecot.index and maildirsize like stuff.
I also thought about just copying everything from one server to other, after checking with one account that it works :D

EDIT:
Based on this topic- https://forum.zentyal.org/index.php?topic=7927.0
Quote
If Zarafa is not used, then MDA relies on Procmail to deliver mails to /var/vmail/domain/user folder.
I think we are using Roundcube...

3
Email and Groupware / Re: Postgres process hogging HDD 90% of the time.
« on: January 02, 2017, 12:36:51 pm »
The machine is "as physical as it can be" - its a separate machine. Backups are made... on the same drive- well I theoretically could plug in some usb drive, then using gui copy everything from that backup location, as we only need emails from it, everything else can go.
The thing is, would I be able to restore those mails and email accounts later on?

4
Email and Groupware / Re: Postgres process hogging HDD 90% of the time.
« on: December 29, 2016, 11:11:07 am »
The reason of not updating is the relic...ness of the server - I have absolutely no warranty that it's gonna reboot succesfully and won't go FUBAR, making my life a living hell since we are using it as our main mailserver. I'd love to upgrade it to at least 4.x, but I don't know if it won't crash it or kill it.
We use mailserver (pop3/imap and smtp) + mailfilter + webmail and ftp- that would be all

5
Email and Groupware / Re: Postgres process hogging HDD 90% of the time.
« on: December 10, 2016, 05:46:57 pm »
Its zentyal 2.2 and it would seem it does something with zentyal components (ebox):

On the other hand, I see deadlocks related to spamassasin's bayesian filter in /var/log/postgresql/postgresql-8.4-main.log.
Still, its hogging the HDD, and I'd love to kill it, but I don't know how it will influence the rest of the services.

6
Email and Groupware / Postgres process hogging HDD 90% of the time.
« on: December 08, 2016, 11:25:06 pm »
Hello, I have the following problem- the postgres process is reading something form disk all the time. Whats interesting, in iotop it has the COMMAND
postgres: ebox eboxlogs [local] SELECT
So I thought its somehow connected to the log module, which is also very laggy for me- so laggy it causes timeouts on the webinterface almost every time I try to do something. I tried disabling the log module, but the process is still hogging the HDD. I'd rather avoid restarting the postgres database, but its really eating the drive...
Do you guys have any ideas on what can I do to lower the disk usage by that process?

7
Email and Groupware / Re: Anti-Spam configurations
« on: December 08, 2016, 10:31:57 pm »
You gotta check the BAYES_00 rule somewhere in the /etc/spamassasin/local.cf
At least that's what I'd do, and it would normally work. The thing is that even if I edit the local.cf file manually it won't start working, cause my zentyal seems to be ignoring it.

8
Hello, I have the following problem-
I have a lot of spam going in from @*.xyz TLD. Since I can't simply put @xyz in the blacklist (and anything with @.xyz or @*xyz can't be typed in) i decided to simply add the line to the /etc/spamassassin/local.cf:
blacklist_from *.xyz
According to SpamAssassin's manual, it should work.
I tried rebooting the mailfilter module, but it got overwritten and i had to redo it.
After I replaced it back with the modified version and did
Code: [Select]
spamassassin -t < [path_to_message_with_now_whitelisted_mail]in console, it worked (-100 score), but the mail was already in mailbox, and in the logs it had 3.6 score.
Now here's the deal - how do I reload the spamassassin/amavis module so that new file won't get overwritten AND will be in use?

[UPDATE]
I've edited the file /usr/share/ebox/stubs/mailfilter/local.cf.mas and added shortcircuiting for blacklisted emails (which was one of the things I also wanted to do) and it worked after restarting mailfilter service.
BUT
I think its a bad idea to tinker with those .mas files AND it takes whitelist/blacklist addresses from some unknown place. For example, I'd like to manually do whitelist_from_rcvd for some mails, but I can't do it manually.

9
So we have the reboot problem solved, but i still have to find out if I can safely kill the lxdm, or just do "Logout" on server terminal (I mean, the monitor that's connected to the PC) without having to do a full reboot.

10
Hello,
We have a mail server which uses Zentyal in 2.0.23. I received the passwords from guy who was going to retire, and basically nothing more, so forgive me my mistakes.
The server is running on Ubuntu 10.04 LTS, and from what I see, its overloaded most of the time (right now it has 3.23 long-term load, even though nothing seems to be hogging CPU for most of the time, except for postgres process, but it seems to be doing stuff on mail), so I thought on what I can kill, to make its "life" easier.
My idea is to kill the desktop GUI (windows - LXDE?), cause I login through SSH anyway.

But while googling the information on whether it is safe to do so, I found some guy talking that doing
Code: [Select]
sudo service gdm stopWill cause the display manager to also kill the foreground apps, and I don't want to kill the server.
Is it safe to do so? Or maybe there is another way of dealing with the desktop gui?

The second question is - how would I normally restart the whole server? I mean, it just seems too easy to go into
System ->
Stop/reboot ->
reboot system.
from the Browser interface. Is it the proper way?
Thanks in advance.

Pages: [1]