Re: How to monitor for a curl error

This discussion got me thinking about syslog and a question came up.

IBM added support in SQL to extract syslog style messages a while back, and after reading the tech note, (search for 634791) it does not appear to need much on the side of IBM i before the messages get dealt with in PACE, where curl would not have the connection issue. There’s a syslogd daemon that will then get the messages over to Greylog where the messages can be handled, ( or ignored) so I’m wondering why the curl from ILE is being used. Furthermore if I’m not mistaken the ccsid needs to be 1200 for the syslog not 37 ( US anyway) so that leaves more questions.

Interesting topic, but it opens up a bunch of questions. Also Lab Services has a product to do this as well, if there is a strong need for it.

Jim Oberholtzer
Agile Technology Architects

On Sep 8, 2021, at 4:42 AM, Patrik Schindler <poc@xxxxxxxxxx> wrote:

Hello Don,

Am 08.09.2021 um 09:55 schrieb Don Brown via MIDRANGE-L <midrange-l@xxxxxxxxxxxxxxxxxx>:

This works but I have found if there are many messages to be written I can
get an error
curl: (7) Failed to connect ... connection refused

This message appears in the joblog

Yes, this is clearly the other side not accepting (another) connection.

My call looks like this;

callp(e) QCmdExc(Cmd:CmdLen);
if %error();

I do get an error but how do I retrieve the actual error that is being sent to the job log (or the qsh session if run interactively) ?

In Unix, shell commands give return values. Sometimes, they're meaningful, sometimes, it's just "if not 0, then error".

After running a shell command, the return value can be used in multiple ways. It can be obtained from the environment variable $?:

true; echo $?
=> 0

false; echo $?
=> 1

(True and false are shell commands just returning 0 or 1.)

So, in theory you can do a

until curl …; sleep 1; done

This makes curl run again and again through a loop until it returns 0. Tried the basic syntax out with 7.2; works.

I have no experience with running UNIX commands in qsh and how/if the return value is passed back to the caller. Maybe others can chime in here?

I think the server is just having a hard time keeping up so if I get this error I was going to retry sending the post or perhaps put a short sleep to handle this issue.

I wonder if hammering each syslog message to Greylog in a separate http request is a good idea at all. :-)

Appreciate any suggestions

I've not yet played with Greylog, so I can't give you an opinion here. Sorry.

:wq! PoC

--
This is the Midrange Systems Technical Discussion (MIDRANGE-L) mailing list
To post a message email: MIDRANGE-L@xxxxxxxxxxxxxxxxxx
To subscribe, unsubscribe, or change list options,
visit: https://lists.midrange.com/mailman/listinfo/midrange-l
or email: MIDRANGE-L-request@xxxxxxxxxxxxxxxxxx
Before posting, please take a moment to review the archives
at https://archive.midrange.com/midrange-l.

Please contact support@xxxxxxxxxxxxxxxxxxxx for any subscription related questions.

Help support midrange.com by shopping at amazon.com with our affiliate link: https://amazon.midrange.com