[1545] in Moira
Re: gen script changes
daemon@ATHENA.MIT.EDU (Garry Zacheiss)
Mon May 8 00:43:22 2000
Message-Id: <200005080443.AAA13009@bobbi-harlow.mit.edu>
To: Jonathon Weiss <jweiss@MIT.EDU>
cc: Garry Zacheiss <zacheiss@MIT.EDU>, moiradev@MIT.EDU
In-Reply-To: Your message of "Sat, 06 May 2000 22:29:01 EDT."
<200005070229.WAA17482@Bearing-An-Hourglass.mit.edu>
Date: Mon, 08 May 2000 00:43:18 -0400
From: Garry Zacheiss <zacheiss@MIT.EDU>
>> The changes you suggest to log the output of the DCM scripts has teh
>> problem that on hosts that get more than one DCM (eg, moira) only the
>> output from the last DCM would be saved. Would it be better to log
>> things to /var/athena/moira_<dcmname>_update.log?
Well, in the specific case you mention, it doesn't matter, because
moira doesn't receieve any dcms that don't use null.sh as their script.
I think the general case of multiple machines receiving dcms is rare
enough that I'm inclined to not care.
>> Logging things to /tmp is also vulnerable to a symlink attack
Wait, we already had this argument. Everyone go back and read the
thread that begins at moira[1353]. I like Dan's reasoning that anything
that doesn't have a /var/athena at this point also doesn't have mortals
logging into it, and putting the output someplace is better than not.
/tmp is only being used if /var/athena doesn't exist, in all cases.
>> Sould we move this to /var/athena too?
I like using /var/tmp for this sort of thing still; if I do decide to
change it, I'd want to do it consistantly across all the gen scripts,
anyway.
>> You need to quote the . from egrep, since . matches any character.
>> You may also want to supress the output of the grep since you don't
>> want to see the hostname, just make sure it is an FQDN.
Sure.
>> The old script rm'd this directory if it existed, you probably still
>> want to do that.
Sure.
>> > ! mv -f $i /var/tmp
>>
>> /var/athena (yeah, I know the old script used /var/tmp.)
>>
See previous comment about preferring /var/tmp to /var/athena for
now.
Garry
