chip@vector.dallas.tx.us (Chip Rosenthal) (05/16/89)
telecom@eecs.nwu.edu (TELECOM Moderator) writes: >X-TELECOM-Digest: volume 9, issue 156, message 4 of 4 >Since last week, copies of the Digest have not been making it through to the >Usenet gateway for some reason. [...] My own control copies of the Digest are >only sporadically making it through to my mailbox at chinet [...] Hi Patrick. Normally I only bother comp.dcom.telecom readers with USENET administrivia, but I would guess that the TELECOM Digest has been really impacted (is that a verb?) by the problems the last few weeks, and the mailing list readers must be wondering too. So, I pass along the gruesome details to them too. The short story is that there were some big problems here at vector.dallas.tx.us. These problems have been resolved, and based on the last week of traffic the link is now reliable. However, one of the fixes requires manual intervention in the gateway processing, and therefore, will add about a day to the turnaround time. The long story is that there were two problems here. The first was that until two weeks ago, vector.dallas.tx.us was running a version of XENIX with a crufty version of uucp. This version had a tendency to throw away messages if the line disconnected, which happens frequently here. The solution has been to: (1) upgrade to a newer version of XENIX with HDB uucp which doesn't throw things away, and (2) install a Trailblazer modem which does a much better job with bad lines. The second problem was that the program which automatically drops the mailed digest into the gatewaying program started to instead drop them into a black hole. The resolution here was to drop them into my mailbox instead, and manually process them from there. Unfortunately, this means that manual intervention is required, and adds some delay to the processing. But I keep traditional hacker's hours, so the delay is usually well under a day. (This should be resolved in 2-3 weeks, when I upgrade to a new mailer here.) Anyway, things appear to be running once again. I appologize for the problems in the interim, and I appreciate yours and Patrick's patience. So, what do you say we all go over to comp.lang.c, kick some butt, and tell them what "#" is really called. :-) (If you missed it, the pound, number, square, octothorpe, tic-tac-toe argument has been going on there for the past two weeks.) -- Chip Rosenthal / chip@vector.Dallas.TX.US / Dallas Semiconductor / 214-450-5337