I think we need to improve trouble shooting. We can send abort in cleanup
scripts. This has been a pending item for years. Why not just do it?
Amf should probably log more and better etc
Skickat från min Sony Xperia™-smartphone
Anders Bjornerstedt skrev
I was mainly thinking about i
I think we need to improve trouble shooting. We can send abort in cleanup
scripts. This has been a pending item for years. Why not just do it?
Amf should probably log more and better etc
Skickat från min Sony Xperia™-smartphone
Anders Bjornerstedt skrev
I was mainly thinking about i
---
** [tickets:#730] AMF: Allow trace-on in healtcheck-callback-reply to increase
timeout**
**Status:** unassigned
**Created:** Fri Jan 17, 2014 02:51 PM UTC by Anders Bjornerstedt
**Last Updated:** Fri Jan 17, 2014 02:51 PM UTC
**Owner:** nobody
When trace is turned on for an OpenSAF serve
- **status**: accepted --> review
---
** [tickets:#690] Opensaf start failed when RDE could not RESPAWN.**
**Status:** review
**Created:** Tue Dec 24, 2013 09:21 AM UTC by manu
**Last Updated:** Fri Jan 17, 2014 12:23 PM UTC
**Owner:** Anders Widell
Changeset:- 4733
Opensaf is up and running.
Has this test been tried without trace on for osafimmnd ?
Having trace turned on will be a major factor in slowing down the
removal of the 200k object ccb from ram in osafimmnd.
If you have note tested without trace and it is not too time consuming to
rerun, then please test without trace. Trace
- **status**: review --> fixed
---
** [tickets:#726] IMM: wrong error codes in README.NO_DANGLING**
**Status:** fixed
**Created:** Thu Jan 16, 2014 12:14 PM UTC by Zoran Milinkovic
**Last Updated:** Thu Jan 16, 2014 12:27 PM UTC
**Owner:** Zoran Milinkovic
In two places in README.NO_DANGLING,
opensaf-4.4.x:
changeset: 4817:55cdaee9338b
branch: opensaf-4.4.x
parent: 4815:c8cb0beaa27e
user:Zoran Milinkovic
date:Thu Jan 16 13:20:37 2014 +0100
summary: IMM: fix wrong error codes in README.NO_DANGLING [#726]
-
default(4.5):
changeset: 4818:4ba0f7b4a
- **status**: review --> fixed
---
** [tickets:#636] IMM: Missing immomtest cases for NO_DANGLING **
**Status:** fixed
**Created:** Fri Nov 22, 2013 09:50 AM UTC by Anders Bjornerstedt
**Last Updated:** Thu Jan 16, 2014 11:58 AM UTC
**Owner:** Zoran Milinkovic
This is related to ticket (#49)
opensaf-4.4.x:
changeset: 4815:c8cb0beaa27e
branch: opensaf-4.4.x
parent: 4807:bb5a37c82405
user:Zoran Milinkovic
date:Thu Jan 16 12:48:29 2014 +0100
summary: IMMTEST: add new test cases for NO_DANGLING flag [#636]
-
default(4.5):
changeset: 4816:2712ca214
Here you are trying to set the scheduling policy to 3 (SCHED_BATCH), which is
not a valid scheduling policy for threads. We can add a check that falls back
to the default policy whenever an invalid policy is encountered.
---
** [tickets:#690] Opensaf start failed when RDE could not RESPAWN.**
- **status**: unassigned --> accepted
- **assigned_to**: Anders Widell
- **Component**: rde --> base
- **Priority**: major --> minor
- **Milestone**: 4.4.RC1 --> 4.2.5
---
** [tickets:#690] Opensaf start failed when RDE could not RESPAWN.**
**Status:** accepted
**Created:** Tue Dec 24, 2013 09
:-) Well yeah, but it atleast gives ideas for alternative approaches!
For eg:- One vague idea could be that FM could check for the existince for any
other core opensaf service when AMFND down arrived and can thus differentiate
an AMFND kill and an opensafd stop, this way it will postpone acting
There seems even a risk that the old active IMMD could beexecuting overlapped
with the new active IMMD executing.
Nooo.
:-)
/AndersBj
From: Anders Bjornerstedt [mailto:ander...@users.sf.net]
Sent: den 17 januari 2014 11:42
To: opensaf-tickets@lists.sourceforg
There seems even a risk that the old active IMMD could beexecuting overlapped
with the new active IMMD executing.
Nooo.
:-)
/AndersBj
From: Anders Bjornerstedt [mailto:ander...@users.sf.net]
Sent: den 17 januari 2014 11:42
To: opensaf-tickets@lists.sourceforg
Hi Mathi,
In this test case it is the amfnd (==AVND I assume) that is killed.
So it goes down first.
That explains why this particular test is so provocative then.
The problem with removing this assert is that it is like opening a can of worms.
It means that I starting to try to support a use-cas
- **Component**: rde --> osaf
---
** [tickets:#721] IMMD asserted when trying to become active during failover**
**Status:** unassigned
**Created:** Thu Jan 16, 2014 07:32 AM UTC by Sirisha Alla
**Last Updated:** Fri Jan 17, 2014 08:37 AM UTC
**Owner:** nobody
The issue is seen on changeset 4
Anders,
Some background information.
FM used to start failover processing upon receiving NODE_DOWN event.
>From 4.4 onwards FM subscribes to AVND down events to start failover
>processing.
This is based on the fact that AVND is the last process to exit (barring AMFD)
i.e. after all the componen
I was mainly thinking about it possibly being overloaded in
execution/processing.
/AndersBj
surender khetavath wrote:
> The system was not at all overloaded. The memory available is ~8GB
> cat /proc/meminfo
> MemTotal:7945404 kB
> MemFree: 7614556 kB
> Buffers: 11980 k
This ticket also raises a general issue about the health-check timer
approach of OpenSAF.
The "out of the box" OpenSAF has quite a large number of such timeouts
in its AMF configuration data, expressed as absolute and real-time values.
But OpenSAF in general will execute on many different platform
The system was not at all overloaded. The memory available is ~8GB
cat /proc/meminfo
MemTotal:7945404 kB
MemFree: 7614556 kB
Buffers: 11980 kB
Cached: 127952 kB
SwapCached:0 kB
Also it is physical m/c not a VM.
---
** [tickets:#724] imm: sync wi
- Description has changed:
Diff:
--- old
+++ new
@@ -22,7 +22,15 @@
PL-3 should go for reboot, but it did not happen.
-Note: For 2N model with above configuration, node went for reboot successfully.
+The same scenario with 2N model observed that amf is crashed
+
+-
+
Yes (you mean huge *number* of objects in one CCB) I suppose
that is one possible fix. But to me it is strange that the IMMND
processing takes such a long time. It is just the RAM processing
we are talking about here, not the PBE processing, since it is the
IMMND that submerges in processing and
---
** [tickets:#729] Node did not go for reboot when error reported on npi
component, recommended recovery as node switchover**
**Status:** unassigned
**Created:** Fri Jan 17, 2014 09:06 AM UTC by Sirisha Alla
**Last Updated:** Fri Jan 17, 2014 09:06 AM UTC
**Owner:** nobody
Changeset : 473
when the SC-2 syslog is analyzed the SC-2 IMMND also got healthcheck timeout at
the same time.
Jan 16 21:57:12 SLES-SLOT3 osafamfnd[4443]: NO
'safComp=IMMND,safSu=SC-2,safSg=NoRed,safApp=OpenSAF' faulted due to
'healthCheckcallbackTimeout' : Recovery is 'componentRestart'
so, IMMD does not fin
- **status**: accepted --> review
---
** [tickets:#728] amfd to exit immediately upon receiving local amfnd down
during 'opensaf stop'**
**Status:** review
**Labels:** AMFD exit AMFND down opensafd stop
**Created:** Fri Jan 17, 2014 08:22 AM UTC by Mathi Naickan
**Last Updated:** Fri Jan 17,
Please ignore the patch in
https://sourceforge.net/p/opensaf/mailman/opensaf-devel/thread/patchbomb.1389947750%40dhcp-hyd-scp-5fl-10-176-178-129.in.oracle.com/#msg31856150
Instead consider the patch in
https://sourceforge.net/p/opensaf/mailman/opensaf-devel/thread/patchbomb.1389948855%40dhcp-hyd-
- **Milestone**: future --> 4.4.RC1
---
** [tickets:#724] imm: sync with payload node resulted in controller reboots**
**Status:** accepted
**Created:** Thu Jan 16, 2014 11:22 AM UTC by surender khetavath
**Last Updated:** Fri Jan 17, 2014 08:42 AM UTC
**Owner:** Anders Bjornerstedt
changeset
- **status**: unassigned --> accepted
- **assigned_to**: Anders Bjornerstedt
---
** [tickets:#724] imm: sync with payload node resulted in controller reboots**
**Status:** accepted
**Created:** Thu Jan 16, 2014 11:22 AM UTC by surender khetavath
**Last Updated:** Fri Jan 17, 2014 07:08 AM UTC
Chenging component for this ticket to RDE.
Possibly it should be FM.
RDE at SC2 is changing role to active while SC2 still has contact
over MDS with SC1.
Jan 15 18:24:01 SLES-64BIT-SLOT2 osafrded[2616]: NO rde_rde_set_role: role set
to 1
---
** [tickets:#721] IMMD asserted when trying to bec
- **status**: accepted --> unassigned
- **assigned_to**: Anders Bjornerstedt --> nobody
- **Component**: imm --> rde
---
** [tickets:#721] IMMD asserted when trying to become active during failover**
**Status:** unassigned
**Created:** Thu Jan 16, 2014 07:32 AM UTC by Sirisha Alla
**Last Upd
---
** [tickets:#728] amfd to exit immediately upon receiving local amfnd down
during 'opensaf stop'**
**Status:** accepted
**Labels:** AMFD exit AMFND down opensafd stop
**Created:** Fri Jan 17, 2014 08:22 AM UTC by Mathi Naickan
**Last Updated:** Fri Jan 17, 2014 08:22 AM UTC
**Owner:** Ma
31 matches
Mail list logo