Closed lichen2013 closed 5 years ago
Tried google, no luck, opened the debug by adding the following line to rsyslog.conf
global(debug.gnutls="10" debug.logFile="/var/log/rsyslogdebug")
And started rsyslog with debug mode: /usr/sbin/rsyslogd -dn
Here is the log file.
upgraded rsyslog version, still error:
Sep 21 02:14:37 qrada-log-forwarder-lbaas-2 rsyslogd[4895]: unexpected GnuTLS error -54 in nsd_gtls.c:594: Error in the pull function. [v8.1910.0.9814b01e74e0 try https://www.rsyslog.com/e/2078 ]
Sep 21 02:14:37 qrada-log-forwarder-lbaas-2 rsyslogd[4895]: netstream session 0x7fe2cc071890 from 135.90.112.13 will be closed due to error [v8.1910.0.9814b01e74e0 try https://www.rsyslog.com/e/2078 ]
rsyslogd 8.1910.0.9814b01e74e0 (aka 2019.10) compiled with:
PLATFORM: x86_64-pc-linux-gnu
PLATFORM (lsb_release -d):
FEATURE_REGEXP: Yes
GSSAPI Kerberos 5 support: No
FEATURE_DEBUG (debug build, slow code): No
32bit Atomic operations supported: Yes
64bit Atomic operations supported: Yes
memory allocator: system default
Runtime Instrumentation (slow code): No
uuid support: Yes
systemd support: Yes
Config file: /etc/rsyslog.conf
PID file: /var/run/rsyslogd.pid
Number of Bits in RainerScript integers: 64
See https://www.rsyslog.com for more information.
Since Version Version 8.32.0 (2018-01-09), there has have a lot of changes in gnutls / openssl code. I suggest that you try latest rsyslog from our repository, to verify if the problem isnt already fixed: https://www.rsyslog.com/ubuntu-repository/
If the problem persists, we can take a deeper look into your problem.
I would assume that this are just connection drops. Gnutls just reports them with a generic error message.
@alorbach I have tried the version 8.1910.0.9814b01e74e0, same issue.
@rgerhards
After several days long run, we have observed same error log at the client side.
Sep 23 21:28:06 lb-bd1247c1-65065 rsyslogd[1495]: unexpected GnuTLS error -53 - this could be caused by a broken connection. GnuTLS reports: Error in the push function. [v8.32.0 try http://www.rsyslog.com/e
Sep 23 21:28:06 lb-bd1247c1-65065 rsyslogd[1495]: omfwd: TCPSendBuf error -2078, destruct TCP Connection to logforwarder.lb.appdomain.cloud:10514 [v8.32.0 try http://www.rsyslog.com/e/2078 ]
When connection drop happens, will rsyslog re-connect and re-send the log that failed to send ? I think we have observed log been miss, and this is a real issue.
read https://rainer.gerhards.net/2008/04/on-unreliability-of-plain-tcp-syslog.html
to be reliable you need to use relp
David Lang
It looks like broken connections.
@lichen2013 you may try openssl ("ossl") driver which provides way better error reporting and handling:
# use ossl netstream driver
$DefaultNetstreamDriver ossl
For more information: https://www.rsyslog.com/doc/v8-stable/concepts/ns_ossl.html
Thanks for the kindly help, @alorbach @rgerhards. Will check relp and openssl. Since these error messages are not real problem, close this issue.
This thread has been automatically locked since there has not been any recent activity after it was closed. Please open a new issue for related bugs.
Expected behavior
Actual behavior
Steps to reproduce the behavior
This is our environment:
log sender (500+ servers with rsyslog configured) ---> log forwarder (2 servers) --> target server
The issue is happening on the log forwarder servers.
After running 2 days, there are a lot of open connections on these 2 servers. It has more than 1 connection from 1 IP. I have checked the specific log sender, only 1 active connection there. So, issue 1 => why there are so many open dead connections on the server side ?
I tried to solve the dead connection issue by adding the keepalive configure:
The connection reduced very fast, but different error shows up in the log:
Environment
servers are VMs on cloud
rsyslog version: