Author: Rainer Gerhards

Isn’t standardization a lovely thing. This morning, I thought I had a new rsyslog release ready. Then I did what was intended to be a quick test on FreeBSD. Guess what – nothing worked at all ;) The reason was that I made the makefile more user-friendly, enabling easier feature selection. For this, I needed to use make file conditionals. As it turned out, they are different between the standard make in Linux and that in FreeBSD. I guess I could have found a make tool for FreeBSD without these incompatibilites – but after all, I want to make it run as easy as possible, which means with as few prerequisites as possible.

I now have split the makefile into a common file and os-specific files in newly created os (or distro) subdirectories. I do not like this approach really much – but as far as I can see currently, it probably is the best. Eventually some Linux wizards will tell me a solution (or I will have a really bright idea), but for the time being, I will have to live with that.

Chris has this morning requested that -syslog-protocol (and the related -transport-udp) be published by the IETF. That’s good news, nothing came up during WG last call. Now eventually the AD come up with something. As I’ve never been through this process, it is quite interesting to me how things progress from now on…

A lot of small development today… For the first time in history (hehe), rsyslog prooves that it is really flexible. We can emulate syslog-ng database writing, so that php-syslog-ng can be used together with it. See the history forum post on using rsyslog with php-syslog-ng. ;)

I am abusing this blog a little as my lab notebook. I had an interesting discussion on IHE and syslog. The issue with that is that IHE defines log records of up to 32K, while syslog only allows records of up to 1k – at least in current standards. Thankfully, many syslog implementation to not take this limit as fixed and ignore the standard in that regard. Also, the upcoming new standard allows for larger messages, so this…

While writing the text, I found blogger to be a bit unhandy for this use. So I moved that over to a paper on our web sites. There it is now, entitled “IHE and syslog message size“.

I keep the text in here as a reference.

Some things iterate from time to time. So this summer’s syslog reliability discussion has surfaced ;) While it for sure is an iteration, it might be slightly different this time. A lot of work has been done on the “reliability front” and much more experience is in the field (and also I have some additional experience and testing done with rsyslog).

Yesterday, I have written a syslog to MySQL tutorial . It describes everything one needs to know to put syslog data into MySQL. Actually, I hope that the description will attract some more people to rsyslog. Let’s see how things evolve…

Actually, I didn’t plan to release a new version of rsyslog today, but it somehow evolved. So I have released rsyslog 0.9.5. It fixes the “semicolon bug” and it also supports multiple rsyslogd instances on a single machine. I needed to support this for our demo system, but it might also be helpful for some secure configurations (I need to think a little bit more about this, but it “smells” like there is a point in this…).

I’ve finally modified rsyslog so that it can run in multiple instances. With that, I could now set up the syslog demo machine so that it runs two instances of rsyslogd. One instance is the “real” rsyslogd, which listens locally. The other instance is the demo rsyslogd, which reads data from the network and shuffles it to the database. As the database was very silent, I were now able to add some rules to forward some events from the real rsyslogd to the demo one. I do this mostly with postfix messages. For demo purposes, I’ve set up a fake postfix. Whoever sends mail to it, gets a bounce back. But the good thing is that postfix has something to do and as such messages will be added to the (demo) system log. I am sure spammers will pick up the mail address from web pages like this one, so I will have a healthy flow of log messages shortly ;)

Finally, I managed to get the new rsyslog site up and running. It turned out to be more work than initially expected. Special thanks go to Timm Herget, who did some of the initial preparation and of course to Andre Lorbach, who made the whole system appear. I just added the content and fumbled a little bit with the config settings ;)

The new site now allows user postings and easier updates. I hope it will be a valuable resource for the (hopefully growing) rsyslog community.

Besides rsyslog, the site also is intended to provide reference for add-ons like phpLogCon.

I so often used ports 1514 and 10514 that I didn’t really notice that they are IANA-reserved. I today posted my syslog stunnel article at aplawrence, and the first comment told me so. Wow! Looks like I need to rethink which ports I used. I am also a bit tempted to try to register a port for syslog/tcp, but this sounds not very easy (especially given the fact that the syslog-sec IETF WG does not at all like plain tcp syslog.