cvs-mirror.mozilla.org is not updating. I haven't seen any new files appear in two days now. cvs stat shows that it thinks I am up-to-date while actually using bonsai to compare file versions the cvs-mirror server is behind.
In the past this was caused by being out of disk space.
I can definately confirm this. Is there any way to pull from the main cvs in the meantime?
I get an access denied from cvs.mozilla.org. Could you check the permissions/allowed hosts, Dawn?
Status: NEW → ASSIGNED
hosts allow seems to be correct. from the rsync log file, here is the last successful rsync from source forge and the most recent failure. They're both from the same IP address so this looks like a dns problem, not a config problem. /etc/rsyncd.conf was last changed sept 26. OTOH, nslookup seemed to behave correctly. The sept 26 change set max connections to 1. Could that be the problem? Also while we're at it could we fix the lines in the conf file that are causing all the error messages? (transfer logging, use chroot, etc) this seems like the problem is on thelizard. reassining to risto. 2000/10/22 19:19:52  rsync on mozilla from sourceforge.net (220.127.116.11) 2000/10/22 19:20:24  wrote 1142229 bytes read 94 bytes total size 930499981 2000/10/25 12:29:46  reverse name lookup failed 2000/10/25 12:29:46  rsync denied on module mozilla from unknown (18.104.22.168) root@thelizard:/etc/> nslookup 22.214.171.124 Server: ns.netscape.com Address: 126.96.36.199 Name: sourceforge.net Address: 188.8.131.52
It looks as though someone has changed the firewall config on mozilla.org's net so that it only lets DNS packets to ns.netscape.com (184.108.40.206) through, and not those to ns2.netscape.com, our backup server, or any other servers for that matter. I've worked around this by commenting out the backup server in /etc/resolv.conf, but this is really bad, since we now have no backup server. Please fix the firewall rules as soon as possible.
Assignee: dtype → rko
Status: ASSIGNED → NEW
Mosedale's change didn't make any difference. the attempted rsync at 15:19 failed again. Note that this firewall horkage is causing other problems as well. We can no longer send mail between thelizard and lounge. Since mail is broken this is keeping people from checking in to cvs.mozilla.org and/or keeping checkins from showing up in bonsai (bonsai.mozilla.org on lounge.mozilla.org). This breakage keeps CPD engineers from working. Marking severity as blocker. Oct 25 15:22:31 thelizard.mozilla.org sendmail: e9PMLRn09456: ruleset=check_mail, email@example.com, relay=localhost [127.0.0.1], reject=451 4.1.8 firstname.lastname@example.org... Sender domain must resolve ntpdate has also been broken since oct 3 Oct 25 14:00:04 thelizard.mozilla.org ntpdate: no server suitable for synchronization found also it looks like weekly dumps have been failing have been failing since sept. USER PID %CPU %MEM SZ RSS TT S START TIME COMMAND root 9842 0.0 0.1 848 704 pts/3 S 15:33:58 0:00 grep vxdump root 2285 0.0 0.1 2080 640 ? S Sep 29 0:00 /usr/sbin/vxdump 6 root 2286 0.0 0.1 2080 640 ? S Sep 29 0:00 /usr/sbin/vxdump 6 root 22611 0.0 0.1 2080 648 ? S Oct 13 0:00 /usr/sbin/vxdump 6 root 22612 0.0 0.1 2080 648 ? S Oct 13 0:00 /usr/sbin/vxdump 6 root 24400 0.0 0.1 2080 640 ? S Oct 06 0:00 /usr/sbin/vxdump 6 root 24402 0.0 0.1 2080 640 ? S Oct 06 0:00 /usr/sbin/vxdump 6 root 27554 0.0 0.1 2080 648 ? S Oct 20 0:00 /usr/sbin/vxdump 6 root 27555 0.0 0.1 2080 648 ? S Oct 20 0:00 /usr/sbin/vxdump 6 I called the help desk and filed ticket 265177 which is being sent to the network team. reassigning to harishm since I think rko is out of town.
Assignee: rko → harishm
Severity: normal → blocker
this turned commercial trunk tinderbox red. zoinks.
adding cls to the cc list.
Dumps... don't worry. If the tape is full it leaves these zombies behind. Usually it fills before the level 6 backups... thus all those leftover processes.
Seems like sourceforge's reverse DNS had been changes, thus *.sourceforge.net didn't match anymore and didn't let rsync. I've added cvs-mirror's IP to the allowed hosts and it seems to sync fine now. Harish, could you verify that this is fixed and if it is please close the ticket.
Running commands from command line seems to be syncing fine to cvs-mirror. Btw, I also raised limits for rsync max hosts. It was set to 1 because that was when we had some preformance problems with MacCVS clients loading the server and in elimination tried that too and forgot to put it back to what it was (3).
Verified with Dawn Endico. And everything except the backup dns server is fixed. There is another ticket alerady open for that.
Status: NEW → RESOLVED
Last Resolved: 18 years ago
Resolution: --- → FIXED
Verified. I just updated my tree and all the changes were retrieved.
Status: RESOLVED → VERIFIED
Product: mozilla.org → mozilla.org Graveyard
You need to log in before you can comment on or make changes to this bug.