From: SourceForge.net <no...@so...> - 2012-09-25 16:26:39
|
Bugs item #3571330, was opened at 2012-09-24 14:06 Message generated for change (Comment added) made by dr_mohan You can respond by visiting: https://sourceforge.net/tracker/?func=detail&atid=532251&aid=3571330&group_id=71730 Please note that this message will contain a full copy of the comment thread, including the initial issue submission, for this request, not just the latest update. Category: HP c-Class Plugin Group: None Status: Open Resolution: None Priority: 5 Private: No Submitted By: Rick Lane (rvlane) Assigned to: dr_mohan (dr_mohan) Summary: openhpid cannot re-connect to OA after network failures Initial Comment: After networking issue that affected openhpid connection to HP c7000 BladeSystem OA, openhpi was never able to reconnect to OA, even after several hours after OA connectivity came back up. Messages in /var/log/messages: Sep 21 05:01:21 HP06-0-0-9 openhpid: ERROR: (oh_ssl.c, 464, BIO_do_connect() failed) Sep 21 05:01:21 HP06-0-0-9 openhpid: ERROR: (oh_ssl.c, 464, BIO_do_connect() failed) Sep 21 05:01:21 HP06-0-0-9 openhpid: ERROR: (oh_ssl.c, 466, SSL error: No route to host) Sep 21 05:01:21 HP06-0-0-9 openhpid: ERROR: (oh_ssl.c, 466, SSL error: No route to host) Sep 21 05:01:34 HP06-0-0-9 openhpid: ERROR: (oh_ssl.c, 464, BIO_do_connect() failed) Sep 21 05:01:34 HP06-0-0-9 openhpid: ERROR: (oh_ssl.c, 464, BIO_do_connect() failed) Restart of openhpid resolved the issue, so openhpid should have been able to resolve the connection issue on its own. ---------------------------------------------------------------------- >Comment By: dr_mohan (dr_mohan) Date: 2012-09-25 09:26 Message: Hemantha, We need to try to reproduce the problem in 3.2.0 with debug on. May be removing the cable or ifdown/ifup of the eth0 could help. If saHpiDiscover succeeds, we need to have good rpt. Did we have similar issues in the recent past. Could you find out? Rick, Are there any other specific steps that we could try? ---------------------------------------------------------------------- Comment By: Rick Lane (rvlane) Date: 2012-09-25 08:42 Message: Another comment - Why did it report discovery done when it couldn't connect successfully? At least if discovery failed, I could add a workaround to restart the openhpid if I cannot successfully perform a discovery within 15 minutes. The way the code currently behaves, not sure how I would even work around this issue. ---------------------------------------------------------------------- Comment By: Rick Lane (rvlane) Date: 2012-09-25 07:13 Message: We have 2.16.0 released currently in our product. Testing on 3.2.0 is unrealistic since this is a customer lab and this is not feasible. This issue occurred when the customer was performing OA and switch maintenance activities that involved power cycling, which would result in intermittent network connectivity. However, after all network connection was restored between host and OA, openhpid never recovered. Mohan has looked at initial logs provided by the customer. ---------------------------------------------------------------------- Comment By: Hemantha Beecherla (hemanthreddy) Date: 2012-09-25 03:20 Message: Hi Rick, The openhpi used in this case is 2.16.0. Could you please test it with 3.2.0 and provide the steps to reproduce the problem?. Thanks& Regards, Hemantha Reddy ---------------------------------------------------------------------- You can respond by visiting: https://sourceforge.net/tracker/?func=detail&atid=532251&aid=3571330&group_id=71730 |