Results 1 to 8 of 8

Thread: Production System Profiles

Hybrid View

  1. #1
    Join Date
    Nov 2005
    Posts
    10
    Rep Power
    10

    Default Production System Profiles

    KevinH listed many possible limiting factors for production systems (http://www.zimbra.com/forums/showthread.php?t=368), but did not give any hard data or even general ideas on how much each factor affected service/performance. This is understandable as the interactions can be varied and complex. He also gives the hardware specs for a production system, but these specs are useless without knowing the factors he listed (e.g. how many users? how much use?).

    Could we possibly get a few production system profiles? Hardware plus usage and performance stats? That way we can approximate new systems to some degree. Squid is another project that has way too many factors to easily predict load/performance; and by having a library of example production system profiles they do their best to help sysadmins. The squid people also have a good document outlining general principles for cache design--telling us which factors are the most important in different scenarios, giving tips, illustrating good design, etc.


    E.g. Hypothetical Example:
    ====================
    Hardware/OS--dual Opteron 3100s, 4GB Ram, 6X200GB software RAID5, FC3 with stock SMP kernel, only services running are Zimbra related.

    Usage--2,500 users, 8am--6pm heavy usage by all of them, they all use the AJAX web interface, no POP, very little IMAP, mixed http/https, quota limit of 10GB, average inbox size of 200 messages, average total disk use of 50MB, total messages sent internally 70,000, sent externally 15,000, received externally 30,000, antivirus is ON, antispam is ON, avge # attachments received every day is 1,000, average # spam every day is 15,000.

    Performance--at 2250 users we began to notice a speed drop. We plan to put in another identical server in the zimbra cluster, migrate 500 users off of the old server, and put all new users on it.

  2. #2
    Join Date
    Oct 2005
    Location
    USA, Canada and India
    Posts
    777
    Rep Power
    11

    Default

    Hi..
    Short answer:
    You cannot have these stats without Release Version.

    Long answer:
    With current version anyway not possible as it has lot of monitoring and other stuff running which will slow down and wont give you correct info, plus Zimbra team is cranking lots of enhancements which will overall can affect web interface performance based on usage.
    Lab test wont give you answers anyway as LIVE INTERNET production server use has its own issues for example our users in India has low bandwidth so if they have to send 1 MB email they connect for 15 min, this has affect on RAM and MTA Processes, imagine 1000 of them on 1 box. Postfix is 100 connections per second or amavisd has limit so your limit will go out of the window.
    in Lab test we try to blast thousands of emails but that works on wire speed with no delays and error, but in real life its different
    Every setup needs to do its own stress test close to real life usage. Distributed install where you put postfix (zimbra) as gateway can increase traffic handling almost double

    What I heard on this message board and my exp with current version of zimbra is
    CPU: Dual CPU 1-1.5 Ghz P4
    RAM: 1-1.5 GB
    HD: SCSI with as much as raid and good speed the better
    HD Space: 3 Times Actual Storage (rough guess, due to Indexes and Mysql usage)
    will do good for 1000 users at average use. But if you getting spammed, dictionary attacked and virus attacked then all this is subjective.

    With your setup example I think 2000-2500 users should be ok.but a good system would be multiple machine setup and you wont need these beefy machines and its good on redundancy.

    * Plain simple postfix + pop with no webmail or web interface can take 5-6 times more load and 3 times less space hungry

    More as we start using Zimbra in production , anyone has any more info in their tests then please post.

    Raj S Vrach
    i2k2systems.com
    Last edited by raj; 11-10-2005 at 07:30 AM.

  3. #3
    Join Date
    Aug 2005
    Location
    San Mateo, CA
    Posts
    4,789
    Rep Power
    19

    Default

    Thank Raj. As you mentioned until we get to GA the numbers for perf and sizing are just guesses. We are making lots of changes in the code and many should help speed things up. More smart caching, better thread managment, etc. That said the performance should only get better. We have no plans to make it worse

    So with that in mind if you need to size your system today just be a little conservative and size for something slightly bigger. That way you can handle the load with plenty of head room when we start making perf improvements.

    So uh-nu-bu how about we turn the question around and you tell us what type of user base you are sizing for and we'll try to recomend something. Over time we'll get more profiles and examples but that won't be until we get more real life data.

  4. #4
    Join Date
    Nov 2005
    Posts
    10
    Rep Power
    10

    Default

    Quote Originally Posted by KevinH

    So uh-nu-bu how about we turn the question around and you tell us what type of user base you are sizing for and we'll try to recomend something. Over time we'll get more profiles and examples but that won't be until we get more real life data.
    How about I turn this back around on you (lol), and instead offer three example situations (small, medium, and large) and ask for appropriate setups for them? That way we have three widely ranging profiles right off the bat, and people can guestimate based on these and any general principles you want to throw out (e.g. outgoing internet email is not very resource intensive, incoming is a bit more intensive if you have antivirus on--and even more intensive if you have antispam on; sending internal email takes almost no resources; imap is very resource intensive and takes about 1-10MB of ram on the server per person using it (if you have 100 ppl using imap clients plan on them using 1GB of ram), pop takes very little resources, the ajax web client is the least resource intensive, etc. etc. -- stuff that seems common sense to you, but to us may not be so readily apparent.

    Small Setup
    =========
    # users: 500
    usage pattern: heavy internal email (send 100 email apiece per day with lots of email lists so 1 email may go to 20 people), lots of calendaring (people depend on the calendars for appointments, scheduling meetings, etc.), all using ajax web user interface. Light internet email (send 20 email apiece per day, receive 40 email apiece per day). Antivirus is on, antispam is on. These people each have a pc, and keep an open window on Zimbra at all times.

    Medium Setup
    ==========
    # users: 2500
    usage pattern: heavy internal email (as above), lots of calendaring (as above), all using ajax web user interface. Light internet email (as above). Antivirus is on, antispam is on. These users each have a pc and keep an open window on zimbra at all times. This company uses a dedicated mail gateway, so the Zimbra server/s can hand off all outgoing email and from which they receive all incoming email.

    Large Setup
    =========
    # users: 10000
    usage pattern: as above with medium setup, except with 2 dedicated mail gateways (mx 10 and mx 20).

  5. #5
    Join Date
    Aug 2005
    Location
    San Mateo, CA
    Posts
    4,789
    Rep Power
    19

    Default

    We just don't have enough real world or lab testing done yet to give this type of estimate. We are planning to provide documentation on sizing but that's not going to come for some time. I think you've seen the recomendtions we've made based on our hosted demo hardware, and some initial tests so until we have more real world and lab data we won't be publishing anything else. As we do get data we'll be sure to let you guys know.

  6. #6
    Join Date
    Nov 2005
    Posts
    10
    Rep Power
    10

    Default

    Thanks KevinH! I'll be looking forward to seeing what you suggest.

Similar Threads

  1. 3.0 to 4.5.3 Upgrade failed (mysql error)
    By dealt in forum Installation
    Replies: 35
    Last Post: 03-19-2007, 11:30 PM
  2. domain coexistence
    By marcmac in forum Administrators
    Replies: 14
    Last Post: 06-30-2006, 02:19 PM
  3. Cannot determine services - exiting
    By tawas in forum Installation
    Replies: 7
    Last Post: 04-25-2006, 03:47 AM
  4. Yet another get.DirectContext issue
    By dccpark in forum Installation
    Replies: 5
    Last Post: 03-08-2006, 01:25 PM
  5. system failure: getDirectContext
    By avisser in forum Installation
    Replies: 3
    Last Post: 10-12-2005, 06:32 AM

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •