climateprediction.net home page
Error -108

Error -108

Questions and Answers : Unix/Linux : Error -108
Message board moderation

To post messages, you must log in.

AuthorMessage
Profile astroWX
Volunteer moderator

Send message
Joined: 5 Aug 04
Posts: 1496
Credit: 95,522,203
RAC: 0
Message 13371 - Posted: 12 Jun 2005, 17:57:03 UTC
Last modified: 12 Jun 2005, 18:46:37 UTC

Running from a folder moved from one Linux box to another, there are Trickles in the queue from each of two runs. (Ebox, P4 3.4 SuSE 9.1 machine's CPDN Directory moved to Dbox, a P4 3.0 SuSE 9.0 machine.)
Boinc forum search: No messages found containing '-108'
Hits on the phpBB but none seem to fit.

2005-06-11 14:20:34 [climateprediction.net] make_scheduler_request: -108
2005-06-11 14:20:34 [climateprediction.net] make_scheduler_request: -108
2005-06-11 14:20:34 [climateprediction.net] scheduler init_op_project to http://climateapps2.oucs.ox.ac.uk/cpdnboinc_cgi/cgi failed, error -108
2005-06-11 14:20:34 [climateprediction.net] scheduler init_op_project to http://climateapps2.oucs.ox.ac.uk/cpdnboinc_cgi/cgi failed, error -108
2005-06-11 14:20:34 [climateprediction.net] Deferring communication with project for 44 minutes and 2 seconds
2005-06-11 14:20:34 [climateprediction.net] Deferring communication with project for 44 minutes and 2 seconds

Perplexing. Spent hours with *.xml files (init_data from both Slots, client_state, global_prefs, sched_reply, sched_request) comparing with runs on other machines. Domain_name is "Dbox" but "hostid" is that of Ebox (Changed Dbox to Ebox but, on retry, it was automagically reset to Dbox.). That may cause accounting problems later but it's obvious I wasn't getting to the server.

Found a previous run's Trickle info at the end of sched_request_*.xml. Aware that can't be correct, I emptied the contents of the four following elements -- hoping for the best trying to make contact with empty elements. Retried and got the "No schedulers responded" message. Progress. Fortunately, the fields were filled-in as part of the request process.

{wu}3ih1_200185556{/wu}
{ph}1{/ph}
{ts}21604{/ts}
{cp}62405{/cp}

{/msg_from_host}
{/scheduler_request}
[Edit: Oops, forgot about not using xml tags...!]

2005-06-12 10:31:48 [climateprediction.net] Scheduler RPC to http://climateapps2.oucs.ox.ac.uk/cpdnboinc_cgi/cgi failed
2005-06-12 10:31:48 [climateprediction.net] Scheduler RPC to http://climateapps2.oucs.ox.ac.uk/cpdnboinc_cgi/cgi failed
2005-06-12 10:31:48 [climateprediction.net] No schedulers responded
2005-06-12 10:31:48 [climateprediction.net] No schedulers responded
2005-06-12 10:31:48 [climateprediction.net] Deferring communication with project for 40 minutes and 1 seconds
2005-06-12 10:31:48 [climateprediction.net] Deferring communication with project for 40 minutes and 1 seconds

[Edit2] Oh, what a tangled web we weave -- when we try to deceive the scheduler:
2005-06-12 11:11:49 [climateprediction.net] Sending request to scheduler: http://climateapps2.oucs.ox.ac.uk/cpdnboinc_cgi/cgi
2005-06-12 11:11:50 [climateprediction.net] Scheduler RPC to http://climateapps2.oucs.ox.ac.uk/cpdnboinc_cgi/cgi succeeded
2005-06-12 11:11:50 [climateprediction.net] General preferences have been updated
2005-06-12 11:11:50 [climateprediction.net] Can't parse scheduler reply
2005-06-12 11:11:50 [climateprediction.net] Can't parse scheduler reply
2005-06-12 11:11:50 [climateprediction.net] Deferring communication with project for 3 hours, 25 minutes, and 53 seconds
2005-06-12 11:11:50 [climateprediction.net] Deferring communication with project for 3 hours, 25 minutes, and 53 seconds

Trickles remain in the queue.

I'm at a loss as to what to try next -- short of restoring the mess to Ebox (and why do I believe that it'd be messed-up there, too...).
___________________________________________
ID: 13371 · Report as offensive     Reply Quote
Les Bayliss
Volunteer moderator

Send message
Joined: 5 Sep 04
Posts: 7629
Credit: 24,240,330
RAC: 0
Message 13377 - Posted: 12 Jun 2005, 19:41:37 UTC

I think there is a problem at Oxford. For the last 3 hours I've been getting:

[---] Can't resolve hostname [climateapps2.oucs.ox.ac.uk] (valid name, no data record of requested type)
[climateprediction.net] scheduler init_op_project to http://climateapps2.oucs.ox.ac.uk/cpdnboinc_cgi/cgi failed, error -113

It's Sunday evening there, so nothing will get done until tomorrow morning. Unless there is a public holday for the Queen's birthday, which will mean Tuesday morning.

Les
ID: 13377 · Report as offensive     Reply Quote
Les Bayliss
Volunteer moderator

Send message
Joined: 5 Sep 04
Posts: 7629
Credit: 24,240,330
RAC: 0
Message 13383 - Posted: 12 Jun 2005, 20:54:54 UTC

My Boo-Boo.
I obviously wasn't fully awake. At 5.30am I went to upload an overnight trickle and found that I hadn't disabled net access, so there were lots of red error messages. BUT I wasn't logged in, so the errors were BOINC not getting to the net.

I just 'woke up', (pun intended), to the problem and tried again, with success.

Sorry all.
(Now, where else did I post?)

Les


ID: 13383 · Report as offensive     Reply Quote
Profile astroWX
Volunteer moderator

Send message
Joined: 5 Aug 04
Posts: 1496
Credit: 95,522,203
RAC: 0
Message 13387 - Posted: 12 Jun 2005, 22:02:31 UTC
Last modified: 12 Jun 2005, 22:42:36 UTC

Update on what started as a -108 problem.
2005-06-12 14:37:43 [climateprediction.net] Sending request to scheduler: http://climateapps2.oucs.ox.ac.uk/cpdnboinc_cgi/cgi
2005-06-12 14:37:44 [---] General prefs: from climateprediction.net (last modified 2005-06-11 07:50:16)
2005-06-12 14:37:44 [---] General prefs: using separate prefs for home
2005-06-12 14:37:44 [climateprediction.net] Project prefs: using separate prefs for home

Contact problem solved; eight Trickles uploaded.

Sched_reply_*.xml also had erroneous values which surely caused the failure to parse; see below. Removed the erroneous runIDs and retried. Success. I think I'll treat myself to a Pilsner Urquell; then see what the server did with the Trickles.

At least two extra machines were created in the process and couldn't be merged with the original. There may be a third after this last restart. (Somehow, one of the spurious machines had credits! Curious, as nothing had been uploaded for either. They merged okay, and the "credits" went away.)

Curiouser and curiouser. I think the moral of this tale is: Avoid playing musical machines with Work Units whenever possible. (I did it with WinXP runs in both Classic and boinc with nothing more than the required hoop-jumping to get squared with Windoze Registry. [Credits all went to the original machine.])

{result_ack}
{name}2lg7_200142336_0{/name}
{/result_ack}
{result_ack}
{name}2q5b_200148484_0{/name}
{/result_ack}
{message_ack/}
{/scheduler_reply}

"name" values were removed; restart or upload or the server set it right.

[Edit]
Another machine was created and, this time, all versions of Dbox merged.
"hostid" 180268 received credit for the eight Trickles uploaded after the "fix". The Trickles show in my Account, but not with the machine. Clicking on the "hostid" in the list of ten:
"Couldn't find computer"

The only Trickles listed for the new machine, #180385, are the last two, one for each active run.

Apparently eight Trickles disappeared into the Great Bit Bucket in the sky. May we be forgiven for thinking this accounting system has something in common with "Hogan's goat"?

Well, anyway, the Pilsner was good!

___________________________________________
ID: 13387 · Report as offensive     Reply Quote
Les Bayliss
Volunteer moderator

Send message
Joined: 5 Sep 04
Posts: 7629
Credit: 24,240,330
RAC: 0
Message 13390 - Posted: 12 Jun 2005, 22:34:06 UTC

Everytime I start thinking about trying Linux on my next machine, someone posts a message like this.
I couldn't get a Slackware version to work decades ago, and when I had to use Unix 10-15 years ago at work, I found it difficult, especially the editor.

Maybe if I just shut my eyes and push the start button. Or maybe by then they'll run them selves.

Les

ID: 13390 · Report as offensive     Reply Quote

Questions and Answers : Unix/Linux : Error -108

©2024 climateprediction.net