autoresponder causing errors

[anewell]

Hi there,

We started seeing bounce emails generated when someone emails an account with an autoresponder active today, though it is not the same error as in this issue. We're running CloudLinux 7.9.0 - the error we're getting is the following:

This message was created automatically by mail delivery software.



A message that you sent could not be delivered to one or more of its recipients. This is a permanent error. The following address(es) failed:



pipe to |/usr/local/cpanel/bin/autorespond [redacted] /home0/[redacted]/.autorespond



generated by [redacted]



The following text was generated during the delivery attempt:



------ pipe to |/usr/local/cpanel/bin/autorespond [redacted] /home0/[redacted]/.autorespond



generated by [redacted] ------



Current version kmod-lve api (10005) doesn't support flag --no-cpu-limit.

Has anyone run into this yet or can point me in the right direction to resolve? Deleting the autoresponder fixes the bounce but I'm not keen on having to delete everyone's autoresponders if I can avoid it.

 

[cPRex]

Hey there! Just for fun, could you try rebuilding CageFS on that system? The last line of that output makes me wonder if that is involved somehow.

If that doesn't work, could you check and see if removing a particular user from CageFS stops the issue from happening?

 

[anewell]

Hey there, I followed the procedure here to rebuild but it hasn't had any effect. I disabled CageFS for one user with the issue and tested and it's still sending the bounce email unfortunately.

 

[cPRex]

Thanks for trying that. Do you have any server updates available that show up with the "yum update" command?

 

[anewell]

I did find a kernel update (3.10.0-962.3.2.lve1.5.70.el7.x86_64) available and have installed it, we'll need to schedule a reboot for the server and see if that's made any difference next. I'll update next when I'm able to, thanks.

 

[cPRex]

I'm wondering if it could because that last line of the output mentions "kmod-lve" which is a kernel module.

 

[dexus]

I did find a kernel update (3.10.0-962.3.2.lve1.5.70.el7.x86_64) available and have installed it, we'll need to schedule a reboot for the server and see if that's made any difference next. I'll update next when I'm able to, thanks.

Did kernel update resolved the issue?

 

[Serra]

It doesn't appear that the kernel version makes this better. I'm running the current kernel and still get the message.

 

[mtindor]

I have friends who run a hosting business, and their CL 7 boxes were experiencing this. They opened a ticket with CL, and CL told them to install latest kernel/reboot. The kernel update DID fix the autoresponder issue for them. You must update the kernel and reboot.

 

[mtindor]

Just got a notification from CloudLinux

Bug in cagefs_enter utility leads to autoreponder issues on cPanel servers.

CloudLinux's Status Page - Bug in cagefs_enter utility leads to autoreponder issues on cPanel servers..

cloudlinux.statuspage.io

CloudLinux Incident - Bug in cagefs_enter utility leads to autoreponder issues on cPanel servers. - 28 July 2022

Bug in cagefs_enter utility leads to autoreponder issues on cPanel servers.

New incident: Investigating

cPanel autoresponder does not work with error:
Current version kmod-lve api (10005) doesn't support flag --no-cpu-limit.
We suppose the issue is caused by a bug in cagefs_enter utility, issue currently under investigation.

Known ways to fix:
CL6 - install cl6 hybrid kernel and reboot.
CL7 - update kernel to the latest and reboot.
CL7h - update kernel and kmod-lve package to latest and reboot.
CL8 - update kernel and kmod-lve package to latest and reboot.
Kernelcare doesn't help to avoid kernel update and reboot.
Removing CageFS also should help, if you can afford it.

Time posted

Jul 28, 12:16 UTC

Components affected

CloudLinux OS Components - CageFS

 

[Serra]

I have friends who run a hosting business, and their CL 7 boxes were experiencing this. They opened a ticket with CL, and CL told them to install latest kernel/reboot. The kernel update DID fix the autoresponder issue for them. You must update the kernel and reboot.

I'll reboot again tonight, but I'm showing the current kernel in KernelCare:

Effective Kernel Version
3.10.0-962.3.2.lve1.5.70.el7

 

[mtindor]

I'll reboot again tonight, but I'm showing the current kernel in KernelCare:

Effective Kernel Version
3.10.0-962.3.2.lve1.5.70.el7

Not necessarily the same thing. Even if you were not running the most recent, you may see that message in KernelCare indicating it has been live-patched to the latest. But a reboot is necessary in all cases after an actual kernel update to fix this issue. Just having kernelcare patching is not enough apparently.

 

[Serra]

Not necessarily the same thing. Even if you were not running the most recent, you may see that message in KernelCare indicating it has been live-patched to the latest. But a reboot is necessary in all cases after an actual kernel update to fix this issue. Just having kernelcare patching is not enough apparently.

Thanks. I'll give it a shot tonight.

 

[mtindor]

Bug in cagefs_enter utility leads to autoreponder issues on cPanel servers.

CloudLinux's Status Page - Bug in cagefs_enter utility leads to autoreponder issues on cPanel servers..

cloudlinux.statuspage.io

Bug in cagefs_enter utility leads to autoreponder issues on cPanel servers.


Monitoring

We have released a fix to beta channel, you can apply it with:
yum clean all
yum update lve-wrappers --enablerepo=cloudlinux-updates-testing

We are monitoring the situation to ensure that the new package doesn't bring any new issues, it'll be released to stable once it's confirmed.

Posted 2 hours ago. Jul 29, 2022 - 07:55 UTC

Investigating

cPanel autoresponder does not work with error:
Current version kmod-lve api (10005) doesn't support flag --no-cpu-limit.
We suppose the issue is caused by a bug in cagefs_enter utility, issue currently under investigation.

Known ways to fix:
CL6 - install cl6 hybrid kernel and reboot.
CL7 - update kernel to the latest and reboot.
CL7h - update kernel and kmod-lve package to latest and reboot.
CL8 - update kernel and kmod-lve package to latest and reboot.
Kernelcare doesn't help to avoid kernel update and reboot.
Removing CageFS also should help, if you can afford it.

Posted 22 hours ago. Jul 28, 2022 - 12:16 UTC

 

[anewell]

I finally had a chance to reboot the server last night and can confirm that updating the kernel and rebooting resolved the issue.

 

[ldxsant]

I finally had a chance to reboot the server last night and can confirm that updating the kernel and rebooting resolved the issue.

@anewell, good morning!

Can you inform the updated Kernel version, please. I'm same problem.

 

[anewell]

@anewell, good morning!

Can you inform the updated Kernel version, please. I'm same problem.

Kernel version on our server (output from uname -rv) is currently: 3.10.0-962.3.2.lve1.5.70.el7.x86_64

 

[ldxsant]

Kernel version on our server (output from uname -rv) is currently: 3.10.0-962.3.2.lve1.5.70.el7.x86_64

@anewell

Thanks!

 

[sparek-3]

Not sure if my issue is related - I'm not using CloudLinux. I've been meaning to study this a bit more and make a post here. But I saw this one and thought I'd chime in.

The issue I'm seeing seems to involve email messages that have a [ or ] character(s) in their From or Reply-To headers (or maybe there are other headers that the autoresponder looks for?)

When run directly the error message given is:

Use of uninitialized value in pattern match (m//) at bin/autorespond.pl line 118, <STDIN> line 67.
Use of uninitialized value in lc at bin/autorespond.pl line 123, <STDIN> line 67.

Again, not sure if it's at all related to this thread's specific issue. Didn't really mean to hijack a thread. But was quickly scouring through this forum and saw this and thought I'd mention it.

I'll do some studying on this later, if anybody from cPanel thinks this is a separate issue, let me know, I'll start a new thread.

 

[cPRex]

@sparek-3 - the issue here is specifically related to CloudLinux. We're tracking that on our end with UPS-472, but that's an "upstream" case and the CloudLinux case number has already been mentioned here.

They'll get that resolved and then things will work as intended.