Hi Minh,
Good catch. I will update in V3.
-----Original Message-----
From: Minh Hon Chau <minh.c...@dektech.com.au>
Sent: Tuesday, May 18, 2021 9:44 AM
To: Thang Duc Nguyen <thang.d.ngu...@dektech.com.au>; Thanh Nguyen
<thanh.ngu...@dektech.com.au>
Cc: opensaf-devel@lists.sourceforge.net
Subject: Re: [PATCH 1/1] smf: enhance smf to handle timeout in one step upgrade
[#3262]
Hi Thang,
Please see comment inline.
Thanks
Minh
On 18/5/21 12:37 pm, thang.d.nguyen wrote:
> In one step upgrade, during the lock nodegroup. The timeout can happen
> and it causes the upgrade failed. By retrying if the return code of
> saImmOmAdminOperationInvoke_2() is SA_AIS_ERR_NO_OP, the lock is
> considered as successfully.
> ---
> src/smf/smfd/SmfAdminState.cc | 13 ++++++++++++-
> 1 file changed, 12 insertions(+), 1 deletion(-)
>
> diff --git a/src/smf/smfd/SmfAdminState.cc
> b/src/smf/smfd/SmfAdminState.cc index 90ae093c4..a75bed798 100755
> --- a/src/smf/smfd/SmfAdminState.cc
> +++ b/src/smf/smfd/SmfAdminState.cc
> @@ -930,6 +930,16 @@ bool SmfAdminStateHandler::nodeGroupAdminOperation(
> (imm_rc == SA_AIS_OK && oi_rc == SA_AIS_ERR_TRY_AGAIN)) {
> base::Sleep(base::MillisToTimespec(2000));
> continue;
> + } else if (imm_rc == SA_AIS_ERR_TIMEOUT) {
> + // Retry
> + continue;
> + } else if (imm_rc == SA_AIS_ERR_NO_OP) {
> + // If an admin operation is already performed SA_AIS_ERR_NO_OP
> + // is returned. Treat this as OK, just log it and return
> + // operation success
> + LOG_NO("Admin op [%d] on [%s], return SA_AIS_ERR_NO_OP,"
> + "treated as OK", adminOp, nodeGroupName_s.c_str());
> + break;
> } else if (imm_rc != SA_AIS_OK) {
> LOG_NO(
> "%s: saImmOmAdminOperationInvoke_2 Fail %s", @@ -948,7
> +958,8 @@ bool SmfAdminStateHandler::nodeGroupAdminOperation(
> }
> }
> if (adminOpTimer.is_timeout()) {
> - if ((imm_rc == SA_AIS_OK) && (oi_rc == SA_AIS_OK)) {
> + if (((imm_rc == SA_AIS_OK) && (oi_rc == SA_AIS_OK)) ||
> + (imm_rc == SA_AIS_ERR_NO_OP)) {
> // Timeout is passed but operation is ok. This is Ok
> method_rc = true;
[M] What if adminOpTimer is not timeout, you get the ERR_NO_OP, then method_rc
is not True
> } else if (imm_rc != SA_AIS_OK) {
_______________________________________________
Opensaf-devel mailing list
Opensaf-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/opensaf-devel
