These are the identifiers currently held.
* The UAF is common for all nodes in the cluster.
* The libraries SYS$LIBRARY and SYS$SYSTEM are different on the good and the
bad node, mainly due to the fact that we are running Vms 6.2 on the bad node
and Vms 7.1 on the good. This has however been the situation quite a while
now.
Process rights:
INTERACTIVE
LOCAL
VAXUS
VAXUS_ADM dynamic
SEBCOMMON_ADM dynamic
SEBTRADE_ADM dynamic
SEBTRADE
DBA_TECH
DBCARE
TDD_ADM dynamic
TDD
SEBIS
SEBIS_ADM
GMD_ADM dynamic
GLOBINFO
SEBINFO
CSD
CSD_ADM dynamic
OMARK
OMARK_ADM dynamic
SEBDIST
SEBDIST_ADM dynamic
HOLIDAY_ADM dynamic
HOLIDAY
SEBAC
SEBAC_ADM dynamic
SEBMSG
SEBMSG_ADM dynamic
GLOBINFO_ADM dynamic
SEBINFO_ADM dynamic
IDA
IDA_ADM dynamic
AMA
AMA_ADM dynamic
SEBRDB
VAXUS_ENV
SEBCOMMON
$
> -----Original Message-----
> From: Dan Sugalski [SMTP:[EMAIL PROTECTED]]
> Sent: Friday, June 30, 2000 9:52 PM
> To: Jordan Henderson; '[EMAIL PROTECTED]'
> Subject: RE: Perl hangs on OpenVMS
>
> At 03:44 PM 6/30/00 -0400, Jordan Henderson wrote:
> >That's kind of the point. The system with java is working fine.
> >
> >They may have had to install a number of patches to the system
> >with java, which indicates that the good and the bad system are
> >running with different shareables. I'm not thinking of
> >incompatibilities here, I'm thinking of corrupted shareable
> >images on the bad system. The shareables may not be the same
> >between the systems, so it's difficult to make comparisons here.
>
> Yeah, but the bad system was working and now it's not. If there was a
> corrupt shareable image I expect nearly everything would be tripping over
> it, so it'd show itself reasonably soon.
>
> On the other hand, there's a biggish bug in the vms_image_init code
> that'll
> cause just this problem if the rightslist for a process is too big. This
> all dredged loose some stuff from quite a while ago, and a quick code
> inspection shows this chunk:
>
> while (rlen < rsz) {
> /* We didn't get all the identifiers on the first pass. Allocate a
> * buffer much larger than $GETJPI wants (rsz is size in bytes that
> * were needed to hold all identifiers at time of last call; we'll
> * allocate that many unsigned long ints), and go back and get 'em.
> */
> if (jpilist[1].bufadr != rlst) Safefree(jpilist[1].bufadr);
> jpilist[1].bufadr = New(1320,mask,rsz,unsigned long int);
> jpilist[1].buflen = rsz * sizeof(unsigned long int);
> _ckvmssts(sys$getjpiw(0,NULL,NULL,&jpilist[1],iosb,NULL,NULL));
> _ckvmssts(iosb[0]);
> }
>
> will fail pretty badly, as there's a while checking rsz and rlen, but
> neither are actually altered in the body of the while. Trip over that and,
>
> well... Boom!
>
> I think there's a patch to this in the archives.
>
> Dan
>
> --------------------------------------"it's like this"-------------------
> Dan Sugalski even samurai
> [EMAIL PROTECTED] have teddy bears and even
> teddy bears get drunk
