RPC Errors

[cfowler]

Hello all,

I wonder if anyone has some greater insight on RPC Errors. I see that the Legato Tech Site,

http://www.fullsw.com,

specifies that the RPC error refers to general network config issues. However, I have an odd situation here. I have clients that are all successfull the previous evening, now they all fail. They are all reporting RPC Error, unable to send, or RPC Error, remote system error. No modifications have been made to the existing network configuration that would justify this error. I just wonder if there exist other reasons for encountering the RPC Error messages.

Thanks in advance,

CF

 

[bozey00]

This just happened to me a week ago. I backup 22 clients and the same thing happened. It turned out that my server index was corrupted, so no connections were being made from the server to any of the clients. Make sure your browse policy and retention ploicies are normal. Also, make sure there is enough space to hold your indexes, since they make copies as well.

In order to fully recover the server information, I had to do the disaster recovery MMRECOV. It wasn't that bad. I can forward you the mmrecov procedures if you want.

Bozey00

 

[cfowler]

Thanks a lot for the info, I greatly appreciate it. Yes, if you could forward the info pertaining to the MMRECOV, I would be most interested to view your procedure. I am currently operating in a Windows NT environment. Might make a small difference if you're coming from a UNIX environment.

Thanks in advance,

CF

 

[SimonBenn]

Hi,

Hold your horses here... you shouldn't need to go for a full DR. Run nsrck -F on the server index multiple times until you can run nsrinfo on it reliably.

Then do the same for the client indexes, with troubles. You need to think why you actually need to have a complete server file index. Do you actually want to restore an old client index given that under 5.5 the b-tree structure is difficult ot merge and very prone to corruption?

I run an 800+ client Networker server (yes that is correct) and if we can't repair the index we throw it out and create a new one. If it's really urgent then you can always run scanner on the SSID.

Regards,

Simon.

 

[cfowler]

SimonBenn,

I appreciate your input. This is valuable advice. However, I was only curious to view the details that bozey00 was willing to forward, for the sake of comparison.

In response to your post, not that you requested a response, both the nsrck -F and nsrim -X options were exercised for my scenario, based on the recommendations of Technical Support. They are the experts and I tend to follow their "sage" advice. The nsrck and nsrim, did not completely solve the problem. You should not have to run either command more than once.

In fact, it would be more assistful to know if there are other reasons for the RPC Error genereation. I have discovered the answer to my initial question. There was one client index that became corrupted. This was not the problem though. The problem was with relation to speed and duplexing. There was a power outage, scheduled, and as it turned out, the UPS was configured incorrectly. When the servers rebooted, some of them came up as Auto / Auto. They were originally set to 100Mbps / Full Duplex. This, in fact, is what began the issue(s) with the RPC Errors. Funny thing is, they relate to two different networks, no routing.

Thanks for your post,

CF