imapd malfunction with RHEL 3.0

P

Protollix

Verified User
Joined
Apr 24, 2004
Messages
54
Hello,
We have two servers running DA right now. One is running RH 9.0 and runs flawlessly.

The other machine is running RHEL 3.0. We have had problem after problem with this machine, and it's all related to RHEL. So I suspect this problem is too.

It seems imapd doesn't spit out attachments properly. it's weird though, because if I access it remotely (like from a mail client using IMAP on my PC) it seems to work ok. but if I access it on the server (via a webmail program) it hangs or only returns a little bit of the image, then distortion. I have tried this with three different webmail clients (Squirrelmail, Ube and DWMail). All exhibit the exact same symptoms.

If I try to access the account from DWMails "demo" installs, the attachment image pops up instantly.

So my guess is that something, somewhere is interferring with local imap connections. but what? I turned the firewall off and that didn't help.

Can someone please help me? This is really getting frustrating
 
My testbed server is currently in use for something else so I can't test this right now.

I should be able to test it in another week.

If this thread stays inactive then please remind me in another week and I'll look into it.

Jeff
 
Welp, I have tracked the problem down to iptables.

Stop iptables, and all the Webmail works perfectly.

I have come to the conclusion that IPtables under RHEL is broken. Or at least mine is. I had this *exact* same problem with a tomcat instance, on this same server. Moved it to redhat 9.0 *with the exact same APF script/config* and it has been up for over a month.

At first I thought it was APF, but even using regular iptables with a few of my own rules eventually makes it die (tomcat).

Somehow, iptables is blocking attachments, but only on the "lo" interface. Even when I config the "lo" interface as "trusted" in APF's config.

Unfortunately, I cannot afford to move my clients again to step down to rh9 or try and have this box reformatted/reinstalled. *sigh*

I really do wish I knew wtf was going on with this. I have about given up on RHEL, worst experience ever with a RH product in over 6 years.
 
Do you mean iptables as in whether it's actually in the kernel and turned on?

Or do you mean whether or not you've actually added rules?

If the latter then it's got to be a matter of testing individual rules.

I'll be happy to test this out once I'm ready to rebuild my testbed system, as I've got a new firewall, based on iptables and on APF, in the works.

Jeff
 
Thanks for the attention man, really :)

It's not individual rules. I did some more testing tonight:

I unload iptables via the command rmmod iptables

Everything works fine.

I then load iptables:
insmod iptables

Everything still works fine

I add *one* rule.
It doesn't matter what the rule is. Repeated tests have shown that after flushing the firewall and adding a single rule, it breaks.

I can load the email and it hangs. I then unload iptables and the page/email finishes loading.

Next test:
unload iptables from the kernel

load ipchains into the kernel (insmod ipchains)

just loading this into the kernel causes the same problem.

I have put a ticket into Server Matrix just to document this. I highly doubt they can fix this. This box is pretty much a pos (well, it's nice specs, but has bad ram or mobo or something). I can't upgrade my kernel becaus e it always panics. Even with the up2date kernel and when hand rolling kernels. When I ran 2.6 (I actually got it to boot a time or two) it wouldn't let me use DMA past udma2.. A host of problems with this box.

I have requested a new box and hopefully I can move everyone over with minimal problems. Interchange is going to be a pain the butt, as it was to setup on this machine.

J, thanks for replies man! I haven't heard of anyone else with this problem, so I am going to chalk it up to a bum server.
 
Thanks for the udpate, Sean.

I hope you'll bring this to my attention in a week or so as I'll be happy to do some testing then.

Jeff
 
Just another update. Acquired a new total control box from SM. RHEL 3.0 also

Everything works as expected. I am unable to duplicate this problem on this new machine.

I will try to remember to bump this again in a week or so in case you still wish to test some
 
I appear to be having this problem now.

squirremail just times out on attachments with iptables on. if i do service iptables stop attachments send fine.

using RHEL 3.0 as well and not sure how long this problem has existed and i'm trying to find a fix as to what could be a different fix.
 
i solved this problem by putting my planet ip addresses in the allow_hosts rule for APF.

think that's ok?
 
I think this is finally coming together...

Van, what do you mean by your planet IP addresses?

Do you mean the server's own IP#s?

If so, then it appears that Squirrelmail is attaching those attachments across the network, which is otherwise being blocked.

Of course it's 2:30 am, so I could be wrong :) .

Jeff
 
yep i mean my own public ips.

however i think iptables was broken or acting up as well. I had all kinds of problems with mysql/coldfusion not running or working together, if i shut off iptables it would work fine. After I played with everything for a few hours I just decided to reboot and everything fixed itself.
 
Yes, but we know mysql works across the network.

So my supposition is probably realistic.

Jeff
 
yeah it was just odd i had the exact same config and only a reboot would solve it.

oh well it works now which is great.
 
well this is still doing it again. I can't figure it out, I have to reboot to get it working. coldfusion stops working with mysql unless i do service iptables stop.

same thing with squirrelmail attachments (When trying to add an attachment it works fine, but when you try to send the message with an attachment it gives all kinds of errors, with iptables off it's fine.)

so odd...anyone with any suggestions what-so-ever would be very much appreciated. i can't be rebooting my box daily.
 
Ok,

making a bit more progress, so if anyone has ANY idea, please please post. This is driving me crazy.

1) it happens about every 2 days (2 days after a reboot i needed to do to fix it)

2) it happens during my sysbk nightly backup.

it happens about 20-30 minutes into the nightly backup i have set.

how could the backup screw up iptables or anything else?

anyone have any insight? any ideas what i can try to make this stop!

thanks
 
well it's confirmed. The DA backup script is what is causing the problems, or least triggering them. after about 20-30 minutes of sysbk running this happens. I have had it off the last 7 days and have not seen an issue.

Not sure what is triggering it, but it's something to do with that script.
 
You might try a manual sysbk, watching the log with a tail command, to see what it's working on when the problem starts.

Jeff
 
i just implemented an rsync backup for now anyway.

it's not such a big deal either way now at least i know what the hell was causing it (it drove me nuts for months).

but i will try that one day, especially if DA backup has a restore feature but right now rsync'n my data is not all the much different.
 
Well, sure it is...

You're keeping one backup up-to-date.

Depending on whether you run rsync continually, or only on a schedule, you may not have a back copy if you or someone else accidentally deletes a needed file.

Just as long as you keep that in mind :).

Jeff
 
You must log in or register to reply here.
Back
Top