Unsolved
This post is more than 5 years old
22 Posts
0
1399
Suppressing MNRESN107W alert - "No HBAs discovered on this host"
I have a Windows server that runs our ECC infrastructure which has no installed HBAs. Each day the following alert is issued to the ECC console:
"MNRESN107W No HBAs discovered for this host. Host may not have HBAs installed or the HBA driver(s) may not be supported."
Since I know that the server has no HBAs I do not want to see this alert any longer. I've opened a case but the solutions offered are less than satisfactory - they are:
1) Acknowledge (but don't clear) the alert so that no new alerts will be issued. Yes, no new alerts would be issued but the acknowledged alert would still persist in the ECC console.
2) Disable alert definitions for all Windows hosts. Of course I still want to see other alerts so this isn't an option.
3) Remove the host from the 'apply to' tab in the alert definitions. Again, not an option because I want to see other alerts relating to the host in question.
4) Resolve the issue. There's no issue to resolve; the Windows host has no HBAs installed therefore I do not want to see a daily alert telling me so.
We're running ECC 5.2. I found a hotfix for ECC 6.0 but there is no hotfix for ECC 5.2.
Can anybody think of a way, short of upgrading to ECC 6.0, that I can disable or suppress this specific alert about a specific server?
Thanks,
Walt
"MNRESN107W No HBAs discovered for this host. Host may not have HBAs installed or the HBA driver(s) may not be supported."
Since I know that the server has no HBAs I do not want to see this alert any longer. I've opened a case but the solutions offered are less than satisfactory - they are:
1) Acknowledge (but don't clear) the alert so that no new alerts will be issued. Yes, no new alerts would be issued but the acknowledged alert would still persist in the ECC console.
2) Disable alert definitions for all Windows hosts. Of course I still want to see other alerts so this isn't an option.
3) Remove the host from the 'apply to' tab in the alert definitions. Again, not an option because I want to see other alerts relating to the host in question.
4) Resolve the issue. There's no issue to resolve; the Windows host has no HBAs installed therefore I do not want to see a daily alert telling me so.
We're running ECC 5.2. I found a hotfix for ECC 6.0 but there is no hotfix for ECC 5.2.
Can anybody think of a way, short of upgrading to ECC 6.0, that I can disable or suppress this specific alert about a specific server?
Thanks,
Walt
ArturC
16 Posts
0
January 22nd, 2008 03:00
That's it. It seems to me that you've got all the possible solutions
Linus
Sabey1
35 Posts
0
October 7th, 2008 10:00
MGG_WIDESKY 12:55:05 C P T 4432 (12.78 MB 28.000000K) [discoverAdaptersAndPorts] => Starting a new
- discovery
MGG_WIDESKY 12:55:05 C P T 4432 (12.78 MB) [[API-Call]] => Issuing SymExtHostHbaGet()
MGG_WIDESKY 12:55:05 C P T 4432 (12.83 MB 52.000000K) [[API-Call]] => Issuing SymPerror()
MGG_WIDESKY 12:55:05 C P W 4432 (12.83 MB) [[API-Call]] => SymExtHostHbaGet() returned with code 3595
- (No objects of the selected type were found)
MGG_WIDESKY 12:55:05 C P T 4432 (12.83 MB) [[API-Call]] => 0 adapters discovered by SymExtHostHbaGet
MGG_WIDESKY 12:55:05 C P W 4432 (12.83 MB) [[Initialization]] => MNRESN107W No HBAs discovered for
- this host.
- Host may not have HBAs installed or the HBA driver(s) may not be supported.
So as you see... they didn't fix it at all. The problem is because you install the Solution Enabler and the API return a bad code for the HBA discovery process (Widesky was a marketing name for Solution Enabler).
And for sure, there is no solution, even today, other than removing the host from the alert, which is a strange way to fix the problem.
seamuscoffey
472 Posts
0
October 7th, 2008 12:00
Apologies if you see the thread marked as 'assumed answered' - my previous post (with moderator privileges) seems to be have put it in that state....not sure why that happened....
Regards,
Séamus Coffey
EMC Global Services
seamuscoffey
472 Posts
0
October 7th, 2008 12:00
Unfortunately, I can confirm that there is no other workaround except the ones mentioned in Walt's original post.
This MNRESN107W alert comes under the 'Generic Alert' defintion - this means if you disable the alert or remove the host in question then you lose the other alerts that come under the 'Generic' definition banner. Ideally, the MNRESN107W alert should be an alert all on its own (as opposed to a Generic one) so that when you disable it, then you know that's the only alert that will no longer be fired.
Note that all Generic alerts are logged in the appropriate agent log even when the alert is disabled in the Console.
For now, your best bet is to keep bringing this to Product Management's attention via an enhancement request in Powerlink.
Regards,
Séamus Coffey
EMC Global Services
NimaiSood
55 Posts
0
October 7th, 2008 23:00
SE libraries packaged with 52sp5 are very old and they do not support the functionality required for the current workaround. The SE libraries need to be upgraded. The correct version of the upgraded SE libraries is V6.4.0.17-824. Engineering ticket was raised for linux only, the changes were common and the SE lib was upgraded for all OS.
I am uploading a latest patch for Host agent patch at the following location ftp://ftp.emc.com/outgoing/11111/
Please download this patch and hopefully this will help to fix this issue. If you have any questions please let me know.
Regards
Nimai Sood
EMC Global Services
tazal
59 Posts
0
October 14th, 2008 09:00
but there is no hotfix for ECC 5.2.
Can anybody think of a way, short of upgrading to ECC
6.0, that I can disable or suppress this specific
alert about a specific server?
I don't think upgrading will help, if I remember correctly the hotfix for 6.0 was for windows hosts that really had HBAs but weren't being discovered properly. Clearly this is not your issue.
I have asked about getting this changed for a long time, having an HBA in a host should most definitely not be an error, I don't even feel it should be a warning.
There are lots of reasons to have host agents on boxes that don't have HBAs, including customers who want to report on non SAN attached hosts using STS or collect performance data via controlcenter.
This is not the sort of thing that should still exist several years in to wanting ECC to be an enterprise level management and reporting application.
Nathan_N_Anthon
17 Posts
0
October 16th, 2008 07:00
Another occurrence of this problem that is painful for us is when a virtual host has the ECC Master Agent installed on it and the ESX server that the virtual resides on is also discovered via the VMWARE agent. We are wanting the functionality of the product to realize that the virtual host is tied to an ESX server, so the virtual wouldn't have an HBA and the alert would never be generated. Our environment has hundreds of these alerts.
The only way we get around them is to acknowledge them and sort by multiple columns in the ECC Console Alerts view by holding the CTRL key and clicking the column headings. This puts all acknowledged alerts at the bottom of the alerts view if you sort by the "Acknowledged By" column first.