Plesk 11.x / Linux :: Slow SMTP Response From Qmail
Jun 16, 2015
We are seeing intermittent slow responses from SMTP on a RHEL6 server running Qmail on Plesk 11.5. The response is being measured from a remote Zabbix server.
The response time seems to be slow (>10s) for a period of 2-3 minutes and then returns to normal (<1s). All other services continue to be ok during the period of slowness.
The server_args line in /etc/xinetd.d/smtp_psa already contains "-Rt0" and all the DNS servers in /etc/resolv.conf are resolving properly.
From looking at the maillog file the server was receiving about 35 SMTP connections a minute at the time of the slowdown. We have the server configured to use 2 x RBL's.
I wonder if we are hitting a limit on the maximum amount of SMTP connections. The file /var/qmail/control/concurrencyincoming does not exist so, according to the Qmail manual, there shouldn't be a limit on the number of incoming SMTP connections.
Smtp service (qmail) stops responding on port 25: # time telnet localhost 25 Trying 127.0.0.1... quit quit Connected to localhost.localdomain (127.0.0.1). Escape character is '^]'. quit quit Connection closed by foreign host.
real 4m10.629s user 0m0.000s sys 0m0.002s
After server restart or sometimes apache stop or ixnetd restart its responding for a some random time, and then again it stops to respond. Plesk panel show it as stopped but qmail itself running in memory, and does other its work, it just stops responding at port 25, or responds with a huge delay.
I've tried change it to postfix, reconfigured with mchk, repaired with repair.sh -r, disabled and uninstalled parallels antivirus, antispam, dnsbl, disabled firewall, disabled smtp lock. Checked apache, dns. Enabled submission port which works when 25 port doesnt, but i need working 25 port.
Nothing solves problem, its just stops responding after some random time. There is no errors on maillog.
I think this problem occured after recent plesk microupdate, because i didn't do anything to server configuration in last months.
This article says it might be dnsbl [URL] .... but it disabled(from plesk panel) on my server, maybe there is way to focefully kill any relation to dnsbl?
Plesk info: OS Red Hat Enterprise Linux Server 5.9 (Tikanga) Panel version 11.5.30 Update #50, last updated at May 18, 2015 05:21 PM The system is up-to-date; last checked at May 17, 2015 10:56 PM
update: xinetd restart is definitely brings smtp alive, but it goes off after random period of time (5min ~ couple hours)
I have a dedicated server that I am renting that appears to randomly respond very slowly for some reason (pings are between 90-4000ms with about 75% packet loss).
This has been happening lately (about once a week for the past 4 weeks) and I thought I had originally found the cause of the problem but it appears I haven't.
Originally, when this problem was happening we were getting an abnormally high about of traffic but a reboot always fixed the issue. I've checked every where I can think of but I cannot find anything out of the ordinary.
ight now we're averaging less than 2Mbps in traffic, the CPU load is less than 0.75, the memory usage is at 55% with no swap being used, Apache connections are normal, no active SSH, FTP, or mail connections, I've restarted all services which had no impact, I've checked the access and error logs but can't see anything out of the ordinary.
I have a ticket open with the data center but I'm trying to figure this out on my own to fix whatever is causing this so I can move my data over to a new server we just got setup over the weekend in a local data center.
Limestone's support is extremely slow. I just got a server with them to host VPSes, so I was inquiring about additional IPs in case I needed them.
Their first two responses were within 4-5 hours, which is bearable. However, yesterday, I actually needed to order more IPs and posted a new reply. I have yet to receive a reply after over 12 hours and my client has been left waiting to receive the additional IPs he ordered.
I haven't been with Limestone long, and I haven't had much other experience with their support, but I was unsure what else to do other than post this here and hopefully get a resolution.
I know their servers are unmanaged and pretty cheap, but I think they should still take the time to answer tickets in a decent amount of time.
Has anyone had similar experiences with Limestone?
I recently made a shift to FDC for one of my servers that I use for downloadable content.
I went for a $199 server with a 100mbps shared bandwidth port. Now, the only problem I seem to be having with it is that downloads appear to be, I would say recognized, very slowly by the server before they start, however, once the download starts I get a full speed. This usually happens during peak times.
I'm using about 10/15mbps on the server according to my cPanel statistics. I just wanted to ask here prior to contacting FDC on this, whether this slow server response time has something to do with the bandwidth port, or perhaps the server's RAM, CPU, etc? Given I'm using quite minimal bandwidth on the port, should I consider having them fix this or upgrade the server instead?
And if I had to upgrade, would I need to upgrade the port or the RAM, or both for this?
This is in regard to ongoing stress I've had over this server from DN...
Keep in mind this is a NEW server order; my first with this company.
I was assigned IPs that appear to be on multiple spam block lists. The "biggest" problems are mail being dumped to bulk by yahoo and dumped to bulk or completely black holed by hotmail. It was also blocked by comcast, mail.ru, and anyone who uses SORBS (those are the only ones I KNOW about so far). This helps reinforce the suspicion that bulk mail was being delivered from the IPs at some point.
I Contacted support at 9PM on the 13th and was referred to abuse. Only after complaining again on the 17th, was I told they requested removal from ONE of the blacklists and they would "look into" the others. I replied saying I wanted new IPs or I would request a refund and that I did not feel the service I recently purchased was actually delivered.
Now here is the part that is giving me the feeling this will not end well for me; I no longer see the abuse department ticket when I log into their ticket system.
What If I ordered the server to use as a mail server? Am I out of line for finding a new server unable to deliver mail to the hotmail/yahoo inbox unacceptable?
Telling my customers to wait and maybe someday their mail will be delivered to the major free mail services is not the kind of reputation I want to have. Its feels like I was sold a problem someone else created and am expected to wait an unspecified amount of time to see if it can be resolved. I do not feel its fair to be left hanging while waiting for white listing attempts on IPs I did not soil in the first place.
I am using Qmail for most of my email accounts which is taking up a lot of my Singapore bandwidth which is extremely expensive. Anyway to configure such that POP goes to my Singapore servers, while I setup a pair of servers in USA for my SMTP needs?
About the middle of the month I started to get SMTP service down messages from a service monitor site I had signed up with. Initially I wasn't concerned but I started to get at least one notice a day. I decided to monitor the server a bit more closely so I setup a virtual appliance on my home server to monitor SMTP response times a bit more frequently -- I let the monitoring application run for a few days to gather data and noticed an unusual trend.
SMTP response times would spike dramatically during business hours -- up to almost 19 seconds at some times. However after business hours and on the weekends response times would return to normal.
My mail volume is really low for my VPS so I don't think it is the problem -- I even ran a mail stress test and pushed close to 500 messages through in 30 minutes (more than I normally get in a day) and I didn't see any SMTP response time spikes. I have also set up monitors to watch some other sites I have running on shared servers with other providers and do not see any unusual SMTP response issues with them at all.
system information is below:
Linux OS: CENTOS Kernel Version: 2.6.18-028stab053.4 Hardware Information: VPS (1500MB RAM) Exim Version: Exim version 4.69 #1 built 16-Mar-2009 16:41:00 Spamd Version: SpamAssassin Server version 3.2.4 running on Perl 5.8.8
I am a new user of Plesk 11.5.30 struggling to understand how qmail is configured.
I intend to install ezmlm to work with qmail but the installation guide assumes that the qmail installation is set up per LWQ (Life with qmail - see URL....
For example the following are quotes from LWQ:All of qmail's system configuration files, with the exception of the .qmail files in ~alias, reside in /var/qmail/control.A series of configuration files resides under /var/qmail/users.
The qmail sendmail, which is normally in /var/qmail/bin/sendmail, usually replaces .None of these directories exist. So where are all these configuration entities?
I have issue with qmail server. I want to update ssl certificate and for this i have replaced older servercert.pem with new servercert.pem file in /var / qmail/ control/. and restart service of qmail using qmailctl restart.But after updating this certificate my web browser still showing older certificate details.
I'm using the API to dynamically create new subscriptions. However, I frequently never receive a response back for my request.
I have enabled graceful restart, which has made an improvement, but my requests still fail to return a response every now and then.
This is problematic, because the script needs to know whether the subscription has been successfully created or not. Often times, when the script times out, the subscription was successfully created. But Once and and a while that's not true, and it is causing inconsistencies in my system.
I have built a database driven site with PHP & MySQL. Pages download quicker on some connections than other (obviously), however this is not due to connection speed. The web host server seems to respond very slowly to my boss's network (which is on a proxy server). Could it be that this problem is related to the web host? i.e. could it be that the web host is unequipped to deal quickly with this type of network?
On my home connection (2Mbps), pages often download instantly, however occaisionally I have to wait 10-20 seconds to receive the page from the host.
Finally, could you recommend an excellent web host to use with our site? My boss wants to offer a professional service where server response is quick. We are based in the UK.
I'm using Plesk 11.5.50 CentOS 6.5 64bit with Qmail. I have installed a SSL certificate on mail server "mail.company.tld" and is running successfully with smtp/pop3/imap4 daemon. Every user agent uses "mail.company.tld" for smtp/pop3/imap4. The qmail name is "mail.company.tld" (file me). The server has about 300 domains and 1000 accounts.
Now we want to add a new SSL certificate, called "mail.newcompany.tld", and use it only for certain domains. I would like to know if is possible to use the new SSL "mail.newcompany.tld" for a specific mail domain without using the old SSL "mail.company.tld" that isn't expired yet. From Plesk Panel I haven't found a section for using the SSL for a specific mail domain.
I have a server(godaddy) with plesk. It was all working well till 8 - 10 days ago. I didnt notice it one week later where I started receiving a lot of failed mail notices. Then on investigation, there were more then 50K spam mails in the mail queue and the mails that were supposed to be sent(registration, forgot password) were also held up in the queue. I found the source of the spam and fixed that.
Also I cleared the mail-queue. Now when I try to send out a test message, it still gets held up in the mail queue. But I can send a test mail to the same domain (email@example.com). All new user registration mails are also held up and this is greatly affecting the site.
Domains are simply not showing in the controlpanel. We can see the subscriptions but its empty under the button 'Domains'. Newly created sites as well as migrated sites do not show.
Before we created or migrated a site to the server we changed psa.conf HTTPD_VHOSTS_D /home/httpd/vhosts This is a legacy from our old servers and never has been a problem. I suppose this is not related, but just thought I should mention it.
The second problem with this server is that when we toggle the mailservice in 'Services management' the button reacts but does not change the state of qmail. It just keeps running. What can be wrong with this fresh installation?