I'm running zimbra on OsX, and really like it.
But I've a Daily crash occuring.
to make zimbra work I did trick the symlink of CurrentJDK to point to 1.5.0 to get java1.5 up and running, and wrote a little StartupItems.
I wasn't really sure if the Java trick and 'my' Statrup Items was faulty or if it was the zimbraServer itself that had a bug (I'm sorry but I did think of it).
Well after digging in the logs I think I finally found the faulty stuff, and IT IS NOT BECAUSE OF ZIMBRA, which is a very good news (I did'nt really think it was the faulty anyway.)

I finally found this trouble entry in the log matching the hanging time:
NetInfo connection failed for server 127.0.0.1.

It seems, there is a bug in OsX (at least tiger) with lookupd, that cause the system to hang, wich is excatly what occurs for me.
I found the clue over here:
http://66.249.93.104/search?q=cache:...er+127.0&hl=en

which points to here:
http://www.macosxhints.com/article.p...51013041404296

The system I'm using to test Zimbra is A G4mirrordrive biproc 2x1,25Ghz and with 1,25 GB of RAM, which is quite correct, I would say.
By reading the above thread in OsXhints, well It seems it only happens with bi-proc and "man-made" ipfw rules
if you are using a hand-made IPFW ruleset, you might want to check that you are not using keep-state rules on dual processor Mac. There is a nasty bug in the kernel/IPFW implementation that will make your machine freeze with that setting. (bug #4112652 on Apple's bugreport, flaged as closed, should be fixed in 10.4.3)
which is effectivly what I've done also because I turn firewall on and edited rules for each ports zimbra uses… (And would say that obviously 10.4.3 did NOT solve this…)

Some suggested that it could be a DNS issue, but I really don't think it my case (I could be wrong but DNS lookup are all working fine on all my Boxes either Mac, Vindoze, or Linux(IPCOP). SO I think (hope) I found the culprit and here I post to share. ANYone testing zimbra with bi-processor could please tell if they have the issue, or anyone with single proc could tell if they Have NOT the issue.

Also here:
http://66.249.93.104/search?q=cache:...er+127.0&hl=en
it is said:
""DirectoryService: NetInfo connection failed for server 127.0.0.1/local."

"The solution was to restore the Netinfo database."

NO. the solution is to turn off "Net Info" in the Directory Access program located within /Applications/Utilities folder.
If you are trying to athunticate to a non-existing netinfo daemon in your domain, your going to get problems.

Turning off that option relieves the problems hinted at in this link. Please sirs, try this instead of blowing away your net info database, When I first got my powerbook 12" I had this same problem. I realized later on that I clicked "Net Info" in the Directory Access program, and it was trying to auth to a non-existant server."
So I'm also going to Directory Access and turn off netinfo, to see if it helps.
AS for sure destroying Netinfo db isn't what I'm thinking of…


So to resume it seems I'm additionning the chance to hang my system as I Have :
-A bi-processor
-"man made" ipfw rules
- Netinfo checked in directory Access.

So i would have to test each one separatly like :
Test#1:
Turning off firewall ? (I'm not on Internet so no pb for testing this)
Test#2:
-Use Only ONE processor with the firewall ON
Test#3
-Turn OFF Netinfo in Directory Access
Test#4:
All of this.

! AS I've installed Apple Dev tools, I can Force to use only I processor instead of 2.
It's been some time (and nightly hours spent) to solve this, so I'm going to Test#4 first and see if this solves the pb, and see if tomorrow morning the system is still up and running or hanged up again.

For sure I'll let zimbracommunity know, here on the forums.

Suddenly I'm thinking to pray :-)… just kidding.