script or probe corruption.

In Jobs that have previously run without error, I will sometimes get an error. The error from the Profile will indicate that the probe or scripts or other element could not be installed in the current base.

in these cases, invariably when the script or element is tracked down, and an attempt is made to view this component using the Local Knowledge entry, the GUI will terminate. Subsequent attempts yield the same results.

any attempt to access the content will die a horrible death.

it is frustrating (but I can recover from this error by deleting the element from the GUI, and recreating it with either a new name or the same name and attaching the blob from one of the managed hosts where the script/probe had previously been executed. Of course onemustidentify which blob contains the errant script)( and of course all references to the old script must be replaced with the new script or element.)

something is corrupting the data set. But I don't know where those references are and so it is difficult to figure out what is disrupting the knowledge base content.

any ideas?

[1132 byte] By [ktensmeyera] at [2007-11-26 18:38:16]
# 1
I recall from one of your other posts that you have been copying blobs around your different servers and so on in an effort to maintain the same setup on them all. Can I ask if this behaviour is noted after this has been done?My initial feeling would be that this could be a problem
ForumModeratora at 2007-7-9 6:12:20 > top of Java-index,Administration Tools,Sun Update Connection-Enterprise Forum...
# 2

well no.

Using an existing blob to use as the content of a replacement script/probe is the issue here.

Last Nov. this problem became much more extensive (corrupted script links), as a result we lost everything in one of our channels.

Approve 4 months of work;

Since the backups also had the corruption, support "suggested" restarting the channel from scratch, and capturing the existing content (in the forms of blobs) to recreate the scripts.

we didn't want to use the originals of the scripts, because so much had been rewritten in place inside of the tool. The only good copies were on targeted server in the form of blobs (but blobs don't have recognizable names) so I had to eliminate earlier copies of the same script on each of the servers.

if I can figure out how to keep it from happing, and if it is still a problem in the next release. I'll ask the customer to find an alternative system.

(and the first step in getting things fixed now, is to add things to the forum, work with the developers, and if necessary fill out service requests)

kermit_at_citia at 2007-7-9 6:12:20 > top of Java-index,Administration Tools,Sun Update Connection-Enterprise Forum...
# 3
Was a bug raised on this behaviour from the support case you mentioned?
ForumModeratora at 2007-7-9 6:12:20 > top of Java-index,Administration Tools,Sun Update Connection-Enterprise Forum...
# 4

I'll most likely need to open a service/support incident report.

just yesterday, we lost the channel.

so today, we've reformated the drive. Install the product again and then restore from the last backup. hopefully then I can recover all the content.

it is likely that the two conditions are releated (missing scripts -and- lost channel)

kermit_at_citia at 2007-7-9 6:12:20 > top of Java-index,Administration Tools,Sun Update Connection-Enterprise Forum...