Great Circle Associates Majordomo-Users
(May 1997)
 

Indexed By Date: [Previous] [Next] Indexed By Thread: [Previous] [Next]

Subject: Re: Majordomo eating lists
From: Jason L Tibbitts III <tibbs @ hpc . uh . edu>
Date: 07 May 1997 19:36:16 -0500
To: cazelski @ BayNetworks . COM (Caroline F. Azelski)
Cc: majordomo-users @ GreatCircle . COM
In-reply-to: cazelski@BayNetworks.COM's message of Wed, 7 May 1997 19:36:02 -0400 (EDT)
References: <199705072336.TAA03370@pobox.engeast.BayNetworks.COM>

>>>>> "CFA" == Caroline F Azelski <cazelski@BayNetworks.COM> writes:

CFA> I had this problem whenever the disk that the lists directory is on
CFA> fills up to 100%.

1.93 did that.  I have been completely unable to duplicate that behavior on
1.94.1, and 1.94.3 should be even better.

I have done some horrendous tests for locking errors and list zeroing,
including filling up TMPDIR while processing 1000 simultaneous unsubscribe
requests.  Of course, this proves nothing (especially for any platform
other than Solaris) but it does at least give an indication of stability.
I have never lost a list in over three years of continuous operation
(except for the one I deleted accidentally).

Now, if you're running Majordomo over NFS, forget it.  Dave's NFS-savvy
shlock.pl might help, but in general there is no safe assumption about
locking over NFS.

As for the "General comment," some feel the same about such "General
comments."

 - J<


References:
Indexed By Date Previous: Re: Majordomo eating lists
From: Michael Brennen <mbrennen@fni.com>
Next: Re: Majordomo eating lists
From: "Kendall P. Bullen" <kendall@lists.tax.org>
Indexed By Thread Previous: Re: Majordomo eating lists
From: Michael Brennen <mbrennen@fni.com>
Next: Re: Majordomo eating lists
From: "Kendall P. Bullen" <kendall@lists.tax.org>

Google
 
Search Internet Search www.greatcircle.com