Koozali.org: home of the SME Server

SME 6.0 - local delivery & server admin probs post crash

Offline telseek

  • 2
  • +0/-0
SME 6.0 - local delivery & server admin probs post crash
« on: July 04, 2005, 10:17:59 AM »
Hi Folks,

After a (failed CPU fan induced) kernel panic & power cycle we've lost local mail delivery and external browser functionality to the server admin. The crash recovery appeared minor - a few inodes corrected, the server popped up and initially we thought there were no problems. But new mail didn't arrive and the users are getting tetchy.

The two problems may not be related - but I've been trolling the archives and comparing files/rights/ownwerships/configs to other SME servers we look after and haven't spotted a cause yet.

** local delivery **

qmail-lspawn (I assume) is logging the following to /var/log/qmail/current

@4000000042c8d38c32532fb4 starting delivery 434: msg 1081590 to local alias-localdelivery-a@b.c
@4000000042c8d38c325356c4 status: local 1/10 remote 0/20
@4000000042c8d38c3277bf14 delivery 434: deferral: Unable_to_run_/bin/sh:_access_denied._(#4.3.0)/
@4000000042c8d38c3277de54 status: local 0/10 remote 0/20

Needless to say /bin/sh and bash are fine and normal. But there's no sign of qmail-local getting up to do its work.


** server admin **

The main page comes up fine, but any selection on the left panel menu results in the main panel coming up with the "Internal Server Error" message.

/var/log/http/admin-error is logging....

[Mon Jul  4 18:02:18 2005] [error] (13)Permission denied: exec of /etc/e-smith/web/panels/manager/cgi-bin/pleasewait failed
[Mon Jul  4 18:02:18 2005] [error] [client 127.0.0.1] Premature end of script headers: /etc/e-smith/web/panels/manager/cgi-bin/pleasewait

Again, pleasewait looks normal.

If I access the server mananger from a ssh/lynx session - it works fine. I presume this handles the frames differently ... not sure if knowing that helps?

Any suggestions on what to look for next would be most appreciated.

TIA, Bill
...

Offline telseek

  • 2
  • +0/-0
Unable_to_run_/bin/sh:_access_denied._(#4.3.0)
« Reply #1 on: July 11, 2005, 01:57:17 AM »
Hi Folks,

Has anybody had any good ideas?

With some desperation, I got around this last time by reverting back to an old age disk set that we'd recently replaced.

To add to the mystery, its now happening on this old disk set as well. The suggests that the orinal fault is unlikely to have been caused by a crash at all. Hopefully there have been no unwelcome/unexpected visitors to the server?

@4000000042d1b3a01af0cc14 delivery 14: deferral: Unable_to_run_/bin/sh:_access_denied._(#4.3.0)/
@4000000042d1b3a01af0eb54 status: local 0/10 remote 0/20
@4000000042d1b3c71aca6bdc starting delivery 15: msg 403285 to local alias-localdelivery-aa@bb.cc
@4000000042d1b3c71aca8f04 status: local 1/10 remote 0/20

I'm going back to the backups to see if I can see if/what has changed... but unfortunately the backups don't include the qmail tree where the problem might be?

Any suggestions most welcome.

Rgds, Bill
...