From chemistry-request # - at - # www.ccl.net Fri Mar 26 01:14:59 1999 Received: from ccl.net (atlantis.ccl.net [192.148.249.4]) by www.ccl.net (8.8.3/8.8.6/OSC/CCL 1.0) with ESMTP id BAA01838 Fri, 26 Mar 1999 01:14:59 -0500 (EST) Received: from krakow.ccl.net (krakow.ccl.net [192.148.249.195]) by ccl.net (8.8.6/8.8.6/OSC 1.1) with SMTP id BAA18009; Fri, 26 Mar 1999 01:14:56 -0500 (EST) Date: Fri, 26 Mar 1999 01:14:55 -0500 (EST) From: Jan Labanowski To: chemistry.,at,.www.ccl.net cc: Jan Labanowski , team-: at :-ccl.net Subject: From the Arministrator: Problems with distributing CCL... Message-ID: Dear CCL Subscribers... While some of you may have enjoyed the silence in the ether, I had many inquires about "What is happening? I am not getting CCL...". For many of us the CCL became the daily annoyance, but the one which is quite addictive. Some popular proverbs about children, wifes/husbands, etc., come to mind. In short... The machine which sends the CCL out to the world had a disk problem (the /var partition which has a /var/spool/mqueue for those who know better) was ailing. The UNIX kernel, rather than sending mail, was frantically trying to read the bad spots on the disk surface. The average load on this machine (Spark 10) went up to 3 or more on occasions. Our sendmail.cf switches off the sendmail queue sweep at load 2, and sendmail refuses connections at 3. We upped it to 3 and 4, respectively, and hoped that load will go down at night and at least the sendmail queue will clear, at which time we take the machine down, put a new disk, and start again, before the ailing drive dies a gruesome death and makes the noise of a sanding machine (sounds familiar {;-)}). But you bet, you loose... The drive's health was more frail, and the life support also failed (this is a spring break time, and the gurus are hard to find). And I have a dilemma... I can resend all the messages (I have them archived on several machines). But at the same time, some of them were sent, some of them were partially sent, and some of them may not have been sent at all. When www.ccl.net vel www.ccl.net died, it had 70,000 or so, pieces of mail to deliver in its queue. And since by definition, you do not backup sendmail queues, since they change by minute, it is hard to tell whose message was delivered, to whom, etc. The suggestions are: Authors: Send your messages again, if you posted them between March 20 and 25, and you did not get any responses. If you do not have your own copy, you can look at gopher://www.ccl.net:70/11/archived-messages/99/03 or if it does not work: http://www.ccl.net/cca/archived-messages/99/03 or ftp://www.ccl.net/pub/chemistry/archived-messages/99/03/ and get them from there. If there are not there, your message was not even received and must have been refused at the moment of spurious high load by www.ccl.net machine. note here, EVERYBODY PLEASE READ... =================================== There are many files there with *.rej in the name. These are messages which had been rejected by the CCL distribution scripts for some reason (yes, computers are stupid, and my perl scripts in particular...), and are usually reviewed by my student assistants, and resent. They should end up in the next day (or so), files with no *rej*, i.e., were successfully resent. If message is in the *rej* file, it does not mean that you are on black list, etc... The message scanning is much more involved, some trivial problems with e-mail header may direct messages for human review. However, if they do not appear there, but appear in rej.pro (processed) and they SHOULD in your opinion, be distributed, please BUG ME (jkl at.at ccl.net) directly. The students who review the messages are not chemists. Moreover, I sometimes personally have a problem if the message is appropriate for distribution. Hard cases make bad law, and some cases are hard. On the other hand, you do not see some of the most offensive messages even in *rej* files, and if we had not had these scripts, you would have learned more about freedom of the Internet, hot sites, fantastic new products, how to be rich in a week, etc., which you may not want to, I presume... Silent Majority: Do not get upset when you see messages being posted again. Thank you for your cooperation, and sorry... I am really sorry for the mess... Now, more bad news {:-)}. Very soon, we will be moving the whole operation to other machines and try to add more redundancy, make the Web site a pleasure to access, etc. And there will be hiccups... But since you love CCL and you are even tempered, and you want to end up in heavens, you will not put a contract on my grey hairy head... But if you want to express your frustration, please send mail directly to me and be direct (I have a skin of a rhino, so get it off...). This will help to keep CCL focused. Yours, Jan Jan K. Labanowski | phone: 614-292-9279, FAX: 614-292-7168 Ohio Supercomputer Center | Internet: jkl %! at !% ccl.net 1224 Kinnear Rd, | http://www.ccl.net/chemistry.html Columbus, OH 43212-1163 |