Hi! I fired up a new portage and source mirror which is reachable at gentoo.inode.at through http, ftp and rsync (portage only via rsync) It's connected with 100Mbit (realistic, may depend on backbone saturation though), although shaping to 20 mbit may happen if there're repeated traffic peaks beyond that limit. The mirror is hosted by inode (an austrian ISP, see www.inode.at) and maintained by me and a workmate. We're reachable at mirror@inode.at . Reproducible: Always Steps to Reproduce: 1. point gentoo to new mirror 2. use it 3. be happy Actual Results: The diskusage on the box increased by ~40 gb :) Expected Results: now this's getting boring =) gentoo@mirror:~$ emerge info -su: emerge: command not found gentoo@mirror:~$
I've added 81.223.20.162 to the access list for rsync1.us.gentoo.org. Please update your rsync cron jobs to point to that server and post back here when you've done so. We'll monitor the server for a couple days to make sure things are working fine and then add it to the rotation. Also, can you confirm that ntp is running? Thanks.
Cron job has been modified, should be running fine. ntpdate is already in cron, the clock gets synced once a day against a stratum 2 server, so this should suffice.
Are you seeing any errors? From the 20:50 on the 23rd through the 02:50 on the 24th (both times UTC), a few syncs were missed. Checked (UTC) Timestamp ================================================ 21:50:03 Wed, 23 Apr 2003 21:01:00 +0000 (expected 21:30) 01:20:03 Thu, 24 Apr 2003 00:33:00 +0000 (expected 01:00) 02:20:02 Thu, 24 Apr 2003 01:33:00 +0000 (expected 02:00)
Hrm, yeah, got a few of these errors: Date: Thu, 24 Apr 2003 03:03:45 +0200 From: Cron Daemon <root@mirror.inode.at> To: robe@amd.co.at Subject: Cron <gentoo@mirror> /data/gentoo/sync-portage.sh rsync error: timeout in data send/receive (code 30) at io.c(85) rsync: connection unexpectedly closed (983411 bytes read so far) rsync error: error in rsync protocol data stream (code 12) at io.c(165) rsync: connection unexpectedly closed (983411 bytes read so far) rsync error: error in rsync protocol data stream (code 12) at io.c(165) --- Date: Thu, 24 Apr 2003 04:05:34 +0200 [snip] rsync error: timeout in data send/receive (code 30) at io.c(85) rsync: connection unexpectedly closed (983471 bytes read so far) rsync error: error in rsync protocol data stream (code 12) at io.c(165) --- Date: Thu, 24 Apr 2003 05:05:03 +0200 [snip] rsync error: timeout in data send/receive (code 30) at io.c(85) rsync: connection unexpectedly closed (983471 bytes read so far) rsync error: error in rsync protocol data stream (code 12) at io.c(165) --- Date: Thu, 24 Apr 2003 10:33:51 +0200 [snip] rsync error: timeout in data send/receive (code 30) at io.c(85) rsync: connection unexpectedly closed (983739 bytes read so far) rsync error: error in rsync protocol data stream (code 12) at io.c(165) --- Seems as if the rsync session stalled after the whole file-list was transferred (nearly identical "bytes read" value). This only happened at rsync1.us.gentoo.org, not to the source mirror at ibiblio... strange. I'm using rsync 2.5.6 over here. Any ideas (except hacking in a "while (!rsyncsuccess) do...")?
I'm wondering if perhaps this a symptom of rsync1 not scaling well. Since all of the servers sync every half hour, rsync1 becomes extremely overloaded every half hour. I'd be interested to see if other servers have experienced similar problems...
Moving the portage tree onto a ramdrive and disabling rsync compression support on the server side might increase the performance (if that hasn't happened already) Do you have load graphs of the box? (hotsanic comes to my mind) Maybe you should use your checking script to verify that the other mirrors are also fetching every update correctly, or are all other mirrors uptodate every time?
Do you have any news regarding this?
I started watching timestamps again. From Thu, 01 May 2003 04:31:00 +0000 through Thu, 01 May 2003 19:31:00 +0000, there haven't been any problems.
yeah, I hardly got any error-mails in the last few days... didn't change anything on my side though
Peter? What's up? I'm using this mirror now for about 2 weeks, and it sure is the fastest gentoo mirror I've ever seen. I'd really appreciate if you could raise it's status to "official" now... regards, Philip
Peter isn't the hold-up. I am. We're working on some things related to our DNS infrastructure. As soon as those are resolved (RSN) this mirror will be added. --kurt
We're ready to add this one... one final question. What is the max user limit?
Regarding the source mirror, I've sent instructions to Michael Renner <robe@amd.co.at> regarding accessing our private distfiles mirror. Please update your rsync scripts to point to this mirror and let us know when you've done so.
ftp is at 30 clients for the vhost, rsync: 30, http: 250 (both pooled for all projects). These values will be increased when/if it should be necessary. Current "usage" can be seen at http://mirror.inode.at/hotsanic/. I've updated the scripts and everything's running fine, a bit slow though (~250kbit).
OK, I added the HTTP and FTP source mirrors to http://www.gentoo.org/main/en/mirrors.xml. They will show up the next time the web servers update (:20 past the hour).
81.223.20.162 is now rsync1.at.gentoo.org. I've subscribed <robe@amd.co.at> to our gentoo-mirrors mailing list (low volume). This list is used to discuss mirror-related issues. Thanks for supporting Gentoo.
Please make sure you are synching to rsync.osuosl.org per the emails sent out to gentoo-mirrors@gentoo.org. We have updated our source tree and your mirror was reported as not having been updated. Thanks! Jeffrey
currently updating against osuosl, may take a while (226kb/sec))
