Message boards :
Number crunching :
How to release orphaned MB WUs?
Message board moderation
Author | Message |
---|---|
BMaytum Send message Joined: 3 Apr 99 Posts: 104 Credit: 4,382,041 RAC: 2 |
Last week (or so) one of my systems suffered (yet another) crash while crunching MB WUs on C2Duo + GTX580 vid card (4 at a time, albeit with questionable 290.xx drivers). Anyway, on reboot and relaunch of BOINC Manager, the PC got assigned a new Seti machine ID for some odd reason, thus the 17 WUs in its cache are orphaned. I hate to let them just expire (not wanting to disservice my honorable wingmen), so is there a way I can "cancel" them for an immediate resend of the WUs? Old PC w/ orphaned WUs here: http://setiathome.berkeley.edu/results.php?hostid=5185956 "New" PC is here: http://setiathome.berkeley.edu/show_host_detail.php?hostid=5510858 I've tried to merge the old PC into the new PC's ID, trying both methods: by name; and manually. But the response is a message saying the "two" PCs are not the same -whoohaa! Is there another way (or some trick to force) the merge? [/url] Sabertooth Z77, i7-3770K@4.2GHz, GTX680, W8.1Pro x64 P5N32-E SLI, C2D E8400@3Ghz, GTX580, Win7SP1Pro x64 & PCLinuxOS2015 x64 |
Cruncher-American Send message Joined: 25 Mar 02 Posts: 1513 Credit: 370,893,186 RAC: 340 |
Last week (or so) one of my systems suffered (yet another) crash while crunching MB WUs on C2Duo + GTX580 vid card (4 at a time, albeit with questionable 290.xx drivers). Anyway, on reboot and relaunch of BOINC Manager, the PC got assigned a new Seti machine ID for some odd reason, thus the 17 WUs in its cache are orphaned. I hate to let them just expire (not wanting to disservice my honorable wingmen), so is there a way I can "cancel" them for an immediate resend of the WUs? You can abort them, I believe, in the Tasks tab. |
BMaytum Send message Joined: 3 Apr 99 Posts: 104 Credit: 4,382,041 RAC: 2 |
You can abort them, I believe, in the Tasks tab. jravin: Yes in theory, I "could" abort them IF my PC could connect to the project using it's OLD ID, BUT now my PC attaches to the S@H project using it's NEW ID (so it doesn't know anything about the OLD PC's orphaned Work Unit cache). Is there a temporary hack (eg an edit in some BOINC xml file(s)) to set the PC to use the OLD ID such that I could then abort the Old PC's WUs, and then "restore" the hacked file)s) back to the New PC ID? Sabertooth Z77, i7-3770K@4.2GHz, GTX680, W8.1Pro x64 P5N32-E SLI, C2D E8400@3Ghz, GTX580, Win7SP1Pro x64 & PCLinuxOS2015 x64 |
HAL9000 Send message Joined: 11 Sep 99 Posts: 6534 Credit: 196,805,888 RAC: 57 |
You can abort them, I believe, in the Tasks tab. I would think you would need to enter the old values into your client_state.xml for <host_cpid> & <hostid>, but there may be more requirements. The merge system is very... wonky. Sometimes it lets me merge completely different machine and other time I can't merge the same machine when it spontaneously generates a new ID. SETI@home classic workunits: 93,865 CPU time: 863,447 hours Join the [url=http://tinyurl.com/8y46zvu]BP6/VP6 User Group[ |
Claggy Send message Joined: 5 Jul 99 Posts: 4654 Credit: 47,537,079 RAC: 4 |
You can abort them, I believe, in the Tasks tab. You're also got to edit the <rpc_seqno> number to a couple of figures higher than the 'Number of times client has contacted server' figure from your host details page, otherwise the server will generate a new host id, Claggy |
Michael W.F. Miles Send message Joined: 24 Mar 07 Posts: 268 Credit: 34,410,870 RAC: 0 |
I have something of the similar problem with total work done My present computer is 6339101 and total credit is 7,547,239 But I have had to redo this computer several times My real total is 10,730,405 on team page The user id for that computer ( The same one now )is 8612202 Is there a way to reclaim to show proper totals on the YOUR COMPUTERS page from SETI I tried to merge but no go there This is what I get when I merge Host 6246514 has overlapping lifetime: (10 Jan 2012 - 4 Apr 2012), (11 Nov 2011 - 24 Jan 2012) Host 6416421 has overlapping lifetime: (10 Jan 2012 - 4 Apr 2012), (10 Feb 2012 - 3 Mar 2012) No hosts are eligible for merging with this one. Michael Miles |
BMaytum Send message Joined: 3 Apr 99 Posts: 104 Credit: 4,382,041 RAC: 2 |
You can abort them, I believe, in the Tasks tab. Thanks jravin and Claggy. I tried your suggestions to "trick" BOINC/ S@H project to think I was connecting with OLD Host ID 5185956, but the trick did NOT take: - I first made a backup copy of my current/new host's client_state.xml file; - I replaced that client_state.xml for my NEW host ID 5510858 with a copy of "client_state.old.xml" (dated 6/6/2010) which has the OLD host ID 5185956 and 22-month old <user_create_time>, <user_total_credit>, <host_create_time>, etc., which BTW has the SAME <host_cpid> value as the NEW host (and many other parameters are the same, e.g. <user_email_hash>, <user_create_date>, <cross_project_id>, etc.); - as Claggy suggested, I also set <rpc_seqno> to 23202 = the project-reported figure for the old host ID 5185956 = 23,199 + 3. After starting BOINC with this hacked client_state.xml, it connects to S@H BUT it doesn't list the 15 orphaned MB Work Units assigned to Old ID 5185956 (it lists 0 tasks in cache = that's what New ID 5510858 has assigned at the moment). Moreover, making that connection caused the hacked client_state.xml to be modified: - e.g., the <cpid_time>, <user_total_credit>, <user_expavg_credit>, etc. were re-written to values for the New host; - yet even so, oddly the <hostid>5185956</hostid> corresponding to the Old host ID was NOT changed. So, do I need to "hack" some other XML file(s) so I can "trick" BOINC/ S@H to allow me to connect as the Old host 5185956 (if so, I'll crunch the orphaned WUs to avoid making Wingmen wait by having them re-sent upon due date expiration). Suggestions anyone?? Sabertooth Z77, i7-3770K@4.2GHz, GTX680, W8.1Pro x64 P5N32-E SLI, C2D E8400@3Ghz, GTX580, Win7SP1Pro x64 & PCLinuxOS2015 x64 |
©2024 University of California
SETI@home and Astropulse are funded by grants from the National Science Foundation, NASA, and donations from SETI@home volunteers. AstroPulse is funded in part by the NSF through grant AST-0307956.