> In the paper introducing journald Seems to b...

2011-12-01T22:08:48.476+01:00

> In the paper introducing journald

Seems to be written by the software developers - not admins. I can see it clearly as a software developer myself.

But I have a little bit of extra experience: I was doing H/W bring up for at least half of year. ("H/W bring up" == try to start Linux on fresh piece of hardware just out of the lab.) And let me tell that dump textual dumb nature of the syslog (and the dmesg) was saving us not once.

On the other system we have installed IBM's alternative (what was it name again) with all the bells and whistles, including fast search and so on... but gosh 'cat | grep' worked so much much more faster compared to the amount of typing all the commands one had to do only to get to the messages.

And binary data were pretty useless too. We dumped some raw network packets into the log and had fun time trying to do anything with them. Not much you can really do about them - simple CSV or textualized TLVs worked for us in the end much much better. Because you can see the data immediately with your eyes. And you can dumbly grep them too.

Comments on Rainer's Blog: Serious syslog problems?

> In the paper introducing journald Seems to b...