upcp hangs on updating courier

[nyjimbo]

Running Freebsd 6.2 and WHM 11.15.0 cPanel 11.18.3-C21703 when I run "upgrade to latest version" for CURRENT it rolls
through a bunch of normal looking checks and after "Updating Bandmin" and then "..Done" the little pop-up box
with the rolling progress bar shows "Updating Courier".

However nothing seems to be happening. If I go to the system console I see the:

Cpanel Update (upcp) - Master (perl 5.8.8)
Cpanel Update (upcp) - Slave (perl 5.8.8)

running and then a "<DEFUNCT>" zombie task below, which a "ps axj" shows belongs to one of the above update tasks.

Nothing else happens after that, no matter how long I leave it running. The defunct process doesnt go away and nothing else starts.

If I kill the above everything seems ok but if I check service status many items are gone and now just see :

exim up
Server Load 0.00 (1 cpu)
Memory Used 32.27 %
Swap 0.00 %
Disk /dev/ad2s1d (/backup) 81 %
Disk /dev/ad0s1d (/tmp) 1 %
Disk /dev/ad0s1e (/var) 12 %
Disk /dev/ad0s1f (/usr) 48 %
Disk /dev/ad0s1g (/home) 37 %
Disk /dev/ad0s1a (/) 3 %


The other regular processes, like apache, exim, ftp, etc are running but not showing up in "service status".

Anyone else seeing this today ?

 

[nyjimbo]

Checking the console I see this come up each time I try to re-run upcp from whm:

Apr 4 23:18:07 xxxxxxx stunnel: LOG5[24834:134656000]: Received signal 15; terminating
Apr 4 23:18:07 xxxxxxx stunnel: warning: can't get client address: Bad file descriptor
Apr 4 23:18:07 xxxxxxx stunnel: warning: /etc/hosts.allow, line 23: can't get client address: Bad file descriptor
Apr 4 23:18:07 xxxxxxx last message repeated 5735 times
Apr 4 23:18:07 xxxxxxx stunnel: warning: can't get client address: Bad file descriptor
Apr 4 23:18:07 xxxxxxx stunnel: warning: /etc/hosts.allow, line 23: can't get client address: Bad file descriptor
Apr 4 23:18:07 xxxxxxx last message repeated 88 times
Apr 4 23:18:07 xxxxxxx stunnel: warning: can't get client address: Bad file descriptor
Apr 4 23:18:07 xxxxxxx stunnel: warning: /etc/hosts.allow, line 23: can't get client address: Bad file descriptor
Apr 4 23:18:07 xxxxxxx last message repeated 44 times
Apr 4 23:18:07 xxxxxxx stunnel: warning: can't get client address: Bad file descriptor
Apr 4 23:18:07 xxxxxxx stunnel: warning: /etc/hosts.allow, line 23: can't get client address: Bad file descriptor
Apr 4 23:18:07 xxxxxxx last message repeated 571 times
Apr 4 23:18:07 xxxxxxx stunnel: warning: can't get client address: Bad file descriptor
Apr 4 23:18:07 xxxxxxx stunnel: warning: /etc/hosts.allow, line 23: can't get client address: Bad file descriptor
Apr 4 23:18:07 xxxxxxx last message repeated 30 times
Apr 4 23:18:08 xxxxxxx stunnel: LOG5[31057:134656000]: stunnel 4.22 on i386-unknown-freebsd6.3 with OpenSSL 0.9.7e-p1 25 Oct

2004
Apr 4 23:18:08 xxxxxxx stunnel: LOG5[31057:134656000]: Threading:PTHREAD SSL:ENGINE Sockets:POLL,IPv6 Auth:LIBWRAP
Apr 4 23:18:08 xxxxxxx stunnel: LOG5[31057:134656000]: 5417 clients allowed

It appears the something in the upcp process is cycling the stunnel process or something. What gets me is that if you look the whole process takes about two seconds yet there are thousands of attempts by something in the system to access it. Nothing weird is running so I dont know what this could be.

Also I see it says "Freebsd 6.3" near the bottom of the above output which is weird as we are running Freebsd 6.2 , even WHM at the top right says "FREEBSD 6.2-RELEASE i386 on standard - WHM X v3.1.0"

Any ideas ?

 

[cPanelKenneth]

I suggest opening a ticket. Both behaviors sound wrong.

 

[nyjimbo]

I suggest opening a ticket. Both behaviors sound wrong.

Cpanel tech suggested doing a portupgrade, which we did and then the upcp ran ok. Still dont know what was causing the hang but the upgrade fixed it somewhere.