ok :(
I've fetched work for other active projects for a day or two and shut network activity while waiting for news. /pg

With All due respect to the Moderator; New Problem, this thread goes back to Oct 2011, it might be new for a few but to some of us it appears to be old.

Maybe it is just that the system is too big, too complicated, and that the support staff are overwhelmed by it all. To me at least it starting to appear like a giant game of whack a mole, patch this problem and the tooth paste finds another weak spot in the system.

Some of us have limited bandwidth and small pipes that we have to share with others, I do not know about you, but my last mile on a really good day is 700K down and 100K up, on a typical day it is on the order of 450K down and 60K up, but that is my problem, but when there is only one service in your area you are just wood chips.

It would be better for us out on the edge for you to shut down the receiving services, Post a notice to Boinc, and message to your message board that the service is shut down until the issue is fixed. That way our local Boinc client will see that service fails to connect and it will keep expand the delay time between attempts to upload as it attempts to perform the upload, I believe that is a feature was installed in the clients to keep them from swamping a server when they were restored to service.

As it is now I have to manually suspend/enable the network, in order to keep this zip file upload from sapping all of what little bandwidth I do have.

---

By 'new', I meant since the university network failure of a couple of weeks ago.
Since this thread started, there have been many occasions of upload problems, which have varied in nature.

The current problem seems to be a repeat of a one from a few weeks ago; a 'disk mount' failure in the storage server that's fed by the upload server.
The upload server won't know that there's no storage server until it tries to transfer the data at the end of the upload.
This is being discussed, and hopefully a fix will emerge to give the uploader some feedback before it OKs the client computer to start an upload.

Also, Oxford Uni has just started it's 'Long Vacation', so there's probably only a skeleton staff looking after network problems. And no night shift.

Other changes have also been discussed in recent weeks. These may take a while, and be transparent to users anyway.

---

Backups: Here

Thanks Thyme & Jonathan. I'm running a WinXP box, so decided editing the xml file might be a quicker/easier solution. This started uploading, but seems like I have similar result with failure at 100% (34Mb file takes about 15mins to up/l).
event log as follows:
27/06/2012 22:23:20 | climateprediction.net | Started upload of hadam3p_eu_4j8v_1999_1_007309876_1_13.zip
27/06/2012 22:32:55 | climateprediction.net | [error] Error reported by file upload server: can't write file /storage/cpdn-restarts/incoming/uploader/hadam3p_eu_4j8v_1999_1_007309876_1_13.zip: Input/output error
27/06/2012 22:32:55 | climateprediction.net | Temporarily failed upload of hadam3p_eu_4j8v_1999_1_007309876_1_13.zip: transient upload error
27/06/2012 22:32:55 | climateprediction.net | Backing off 4 hr 0 min 13 sec on upload of hadam3p_eu_4j8v_1999_1_007309876_1_13.zip

Will hang about to see if this clears in the next few days.
Thanks for your guidance.

Steve

I'm afraid that you've come up against another problem that started a few hours ago.
If you can, Suspend the BOINC network connection to save your bandwidth.

---

Backups: Here

yes I figured. Meanwhile I went off and had a look at BAM and it said it wasn't communicating right with CPDN, so *bright idea* I disconnected the project in my manager...and seems like that reset the project...oops. So end of thread here me thinks. Is the upload file still stored and that I can send manually or have i really blooped?
Also is there a discussion somewhere on the BAM/CPDN in these forums that I can contribute to?

Yes, Reset does just that - deletes everything. This is mostly for when a project program has become corrupt, and you need to get a new copy. But all of your data files go too. :(
When I said Suspend, I meant in the BOINC manager's menu.

Re: BAM/cpdn. Not really. There have been posts in the past about problems with this combination, because for some reason, some of the BAM functions don't work with updating cpdn stuff. Changes wanted, e.g. to pref settings, need to be done manually here, and not through BAM.

---

Backups: Here

Upload from full ocean model went through fine - presumably going to a different server and pnws except for the 13 going to Oregon State mean I should be OK for a few days.

_13.zip files from the regional models are uploading OK now.