Unsolved
This post is more than 5 years old
32 Posts
0
1550
nsrindexd Failed to poll() on incoming connections: An operation was attempted on something that is not a socket
Hi
We got a problem backuping Domino server (AIX) with Networker. Sometime (often), backup is hanging during TXNLogs backup, and we got this message on the server side:
nsrindexd Failed to poll() on incoming connections: An operation was attempted on something that is not a socket
Please help (SR is opened but any other idea is welcome)
Thanks
Christophe
coganb
736 Posts
0
March 11th, 2011 06:00
Hi Christophe,
Does this only happen for Domino backups? Does reducing the server parallelism 'solve' the problem? If this is only happening for one group, is that the biggest group? Is it only happening when the server is busiest? When did it start happening?
-Bobby
clepron1
32 Posts
0
March 11th, 2011 07:00
Hi Bobby
yes it seems to happen only on domino AIX servers, especially during
transaction log (TXN) backup (but sometimes on databases backups).
Parallelism is equal 2 .
We got 5 AIX domino servers, with same config, and same issue. Not all the
time, but often (I should say one issue for two backups, in average).
Do you advice to reduce it to 1 ? As the size of each servers is 2 TB, it
could take longer, but we can try if you think so.
It happens on two groups, but could happen when only one server is
backuping.
It was working wel during more than one month (the networker
infrastructure is newly installed, at the end of 2010), without issue. We
began to have problem when we plug the enclosure to the 10Gbits backbone.
(no issue seen on the network, no packet loss for example). The enclosure
has 1Gbits switch, but is not overloaded. (700Mbits in average during
backup)
Networker servers has 10Gbits connectivity.
Thanks for your help .
Christophe
coganb
736 Posts
0
March 11th, 2011 07:00
Christophe,
I meant the server parallelism, not the client parallelism but if this happens even when there is only one group running, that's probably not the issue. You could try slowing things down by modifying the group or client parallelism to see if that changes anything. What's important aswell to find the source of the problem is to know whether this only happens on Domino backups because they are the largest backups or because there is something else specific to them or whether this is really Domino-related.
-Bobby
ble1
2 Intern
2 Intern
•
14.3K Posts
0
March 11th, 2011 18:00
A message like that, but for nsrexecd, has been addressed in 7.6.1.4 and 7.5.4.2 (now available). Not sure if and how that may be related to your issue. Perhaps if running a build close to above ones you could try to update to check if that addresses your issue or not. Since you have case opened, asking support first might be first step.