Hostserver Socket Demons Not Start
Masaki Satoh
I have a ploblem, STRHOSTSVR *ALL doing but socket demons not start.
I did NETSTAT option 3,but demon port was not listen status.
I did WRKACTJOB SBS(QSYSWRK QSERVER),but I couldn't look at every demon.
I did WRKOUTQ QEZJOBLOG,but about hostserver spool was not cumlative.
I did WRKJOBQ QSYSMAX,but DEMON JOB was not entry.
I did WRKSBD SBS(QSYSWRK QSERVER) checkin routing table,but not ploblem.
Everybody PLEASE GIVE ME A HINT !
thanks
Matt Griffith
One of three things (that I know of) is causing your problem. First try the
suggestions from usenet posting five days ago.
Assuming, all the other as-* daemons are running, and assuming you have
followed up
with a STRHOSTSVR *SIGNON and the as-signon daemon still does show up with
a
NETSTAT option 3, you should first run ENDHOSTSVR *ALL (which I doubt will
be a
problem 8-) ) and then run WRKLNK. You should see a 'qserver' link. Rename
qserver
to (saay) qserverold. Then run STRHOSTSVR *ALL and voila.
Please note the last few times I have encountered this problem it has been
on V4
and the WRKLNK path is different. To be sure I specified the correct path
above,
take option 5 next to 'qserver' link and you should see objects which
correspond to
all the other as-* daemons with a ".end" extension. If so, this is the
correct
path.
If anybody has the same problem on V4, the path is
'/QIBM/ProdData/OS400/servers'
(I think).
May not be your case, but a lot of times this problem occurs after recovery
from a
power outage.
If that (RENAME LINK trick) does not apply, I'll bet you are on V4RX. If
so, take a look at *JOBD QZBSJOBD. It's Initial Library List has QGPL and
QTEMP. If either one of these libraries are included in the SYSVAL QSYSLIBL,
you will have host server problems. You will have to edit either the SYSVAL
or the JOBD.
The only other thing you can check (and this is pretty vague), is again,
your SYSVAL for QSYSLIBL. If this value is rather bloated and includes a
large number of entries, try trimming the fat and see if you get better
results.
Masaki Satoh
Thank you Matt your advice.
But this ploblem is not resolved.
I checked WRKLNK , not created qserver.
SO manual CRTDIR qserver and additional authority *public *rwx.
I try again STRHOSTSVR *ALL, but hostsever not started.
AND ENDHOSTSVR *All,but there are not demon-names.end (type socket) under
qserver dir.
I checked *JOBD QZBSJOBD.none ploblem.
I try JOBQ hold QSYSNOMAX , QPWFSEVER ,so QXXXXX demons CHGJOB LOG
*SECLVL.
I saw logs ,so I looked at CPIAD08 "Communication error on hostsrver on
mkdir()-
host sever domain - path".
But I create qsever dir and authority.
Why reason ?
OS400 r370(C8069370) r320(?) same ploblrm
thanks
<<============================
<< Name : Masaki Satoh
<< E-mail : sat...@sb3.so-net.or.jp
<<============================
Matt Griffith wrote in message <35873B49...@texas.net>...
Matt Griffith
Masaki Satoh wrote:
> Thank you Matt your advice.
>
> But this ploblem is not resolved.
>
Okay, lets make sure we are in synch. After ENDHOSTSVR *ALL, and renamed (or
deleted) '/qserver' link, and then ran STRHOSTSVR *ALL, the '/qserver' was still
not there? If so, that is not good. STRHOSTSVR is supposed to recognize the link
is not there and create a new one. Just to be clear, you did not run MKDIR
before giving STRHOSTSVR a chance to create what it needs?
Also, you mentioned message id CPIAD08. The generic text on this message is to
see previous listed messages in the joblog. Are there any previously listed
messages? An old-fashioned authority problem my be the cause.
Have you checked out USRPRF QUSER? Is the USRPRF *ENABLED? Is Set Password to
Expired = *NO? Is Maximum Storage Allowed set to *NOMAX?
Another puzzler, did you actually have a JOBD QZBSJOBD on your system? I'm not
in front of a 400 right now, but I thought this JOBD came into existence with
V4R1. If nothing listed above fixes your problem, I would want to lean heavily
on library list problems. Hopefully, if this is the problem, your joblogs will
have a message with text to the effect "Duplicate Library."
Also not previously mentioned, run CHKPRDOPT for LICPGM 5716SS1 OPTION 12 and
make sure "no errors detected." That's for your V3R7 box - use 5763SS1 on your
V3R2 box.
Also, I cannot provide you any great details right now because I am relying
solely on (bad) memory, but I seem to remember a problem with the **ENDHOSTSVR**
command that was unique to CCSIDs and code pages (I'll confess, I do not fully
comprehend CCSIDs, etc). It seems that as-* daemons were still hanging around
after ENDHOSTSVR *ALL command was run. Are there any as-* daemons showing up
with NETSTAT option 3 after running ENDHOSTSVR? If so, take Option 4 and zap
them and then try STRHOSTSVR *ALL.
If all else fails and no solutions are in sight, try deleting SS1 option 12
(please, option 12 only 8-) ) and reinstall. Be sure to reapply cume pack after
reinstall.
Good Luck.
Cheers, Matt.
P.S. - The fact that you are experiencing this problem on two different machines
leads me to believe that you either have an authority problem or a library list
problem.
Masaki Satoh
Thanks Matt deep regurads.
You taught me an authority ploblem ,and I resoled this ploblem.
For reason root(/) authority *public *exclude.
*public changed *rwx , and strthostsvr *all is running.
I am poor checking how to myself !