scgiwrap: Caller must be uid 99

[jenlepp]

I upgraded to Apache 2.2 this past weekend, and everything looked to be going ok until a client emailed in getting this error on their script.

scgiwrap: Caller must be uid 99

When I had Apache 1.3, the suexec_log looked like this:

[2008-01-19 05:13:56]: info: (target/actual) uid: (blah/blah) gid: (blah/blah) cmd: guestbook.cgi
[2008-01-19 05:13:56]: info: (target/actual) uid: (blah/blah) gid: (blah/blah) cmd: guestbook.cgi
[2008-01-19 05:13:57]: info: (target/actual) uid: (blah2/blah2) gid: (blah2/blah2) cmd: guestbook.cgi
[2008-01-19 05:14:32]: info: (target/actual) uid: (blah2/blah2) gid: (blah2/blah2) cmd: agora.cgi

Now with 2.2 after recompile, they look like this.

[2008-01-20 14:37:20]: uid: (32307/blahblah) gid: (32309/32309) cmd: Count.cgi
[2008-01-20 14:37:23]: uid: (32427/blah) gid: (32429/32429) cmd: randomvideo.pl
[2008-01-20 14:37:23]: uid: (32427/blah) gid: (32429/32429) cmd: currentupdatex.pl
[2008-01-20 14:37:23]: uid: (32427/blah) gid: (32429/32429) cmd: allusers2.pl

and they get the errors above. I've recompiled, and still the same thing - is there some step that I missed somewhere that I should have done but didn't do? This is happening on all servers. Also getting an occasional error in the log of:

[2008-01-20 14:38:43]: cannot stat program: (date --utc)

[jdlightsey]

If you update to the latest release of cPanel/WHM, go into the "Configure PHP and Suexec" page on the WHM, change the Suexec setting and save, then change it back and save, does it fix the problem?

We recently made some changes to the way the suexec binary is managed. If that doesn't sort out your problems I'd be very interested in seeing exactly what is going wrong. Submit a ticket "Attn: JD" if you can.

[jenlepp]

I did try turning it off, and then turning it back on again in the way you suggested, but the script still had the error. I wound up just turning it off completely until I could figure out what was going on.

I get cpanel from my NOC and so they have to submit the ticket - I'll point them to this response and see if we can't figure out what's up.

[cpanelnick]

I did try turning it off, and then turning it back on again in the way you suggested, but the script still had the error. I wound up just turning it off completely until I could figure out what was going on.

I get cpanel from my NOC and so they have to submit the ticket - I'll point them to this response and see if we can't figure out what's up.

You can still submit a ticket at https://tickets.cpanel.net/submit/ The complimentary queue is pretty small right now so it should get sent to jd quickly (he won't be back till tomorrow morning though most likely)

[s.a.]

Here's the reply I got from support regarding this issue:

"It would appear that scgi has been deprecated in newer cPanel versions. I recommend you move these scripts to the cgi-bin."

[cpanelkenneth]

Here's the reply I got from support regarding this issue:

"It would appear that scgi has been deprecated in newer cPanel versions. I recommend you move these scripts to the cgi-bin."

Open a ticket in our complimentary queue, please.

[s.a.]

Open a ticket in our complimentary queue, please.

Where's that? I used tickets.cpanel.net

[cPanelDavidG]

Where's that? I used tickets.cpanel.net

Go to http://tickets.cpanel.net/submit and at the bottom where it says submit a ticket without logging in, follow those instructions .

[s.a.]

Sorry I just don't get it.
Do you mean I might get a different reply than I got from priority support few hours before?

[cPanelDavidG]

Sorry I just don't get it.
Do you mean I might get a different reply than I got from priority support few hours before?

What was the ticket ID number?

[s.a.]

It's 234128.

[cpanelkenneth]

Check the ownership/permissions on the scgi wrapper script/binary. The full path is in httpd.conf.

[jenlepp]

I have done all that. Also worked on it with my NOC. They caught what we didn't.

I'm downgrading back to apache 1.3 - I have way too many broken sites, once got hacked today thanks to turning off suexec, and I'm outta patience.

Thanks.

[jenlepp]

Well, I have fixed it - I downgraded, and the correction in the suexec log was immediate. As much as I'd love to help out while it gets looked at, I can't leave a production server with multiple sites broken for days on end - these are hundreds of clients, not my personal fun sites that can sit there broken while everyone says "hmmmm".

Since it happened on all four servers at the same time when I did the same thing, I'm sure it's not just me, but I think I'll wait for someone else to be the guinea pig before jumping in like that again.

[dflteche]

Hello,


Add following entry in httpd.conf file on the server :

AddHandler cgi-script .cgi .pl

It works for me