Nserver.exe 100% utilization

HCL Domino Domino Forum
1

Since upgrading from 8.02 to 8.5 nserver.exe appears to be stuck in some loop with approximately 100% utilization of one processor indefinitely. Also noted that Kernel utilisation (process name “System”) appears to track closely at about 15% utilization.

Eventual crash (unknown trigger) with callstack:

@[ 1] 0x60001a29 nnotes.Cmovmem@12+329

(b0f0010;f7cb206;743b)

@[ 2] 0x60a1b884 nnotes.DB2SQLFetchNote+2516

(b0ef110;253967c;b0eef1c)

@[ 3] 0x60a144f4 nnotes.ForEachDB2Document+180

(b0ef110;253967c;842a)

@[ 4] 0x60a13dec nnotes.iDB2SearchFill@8+3068

(2d7b;2d5f)

@[ 5] 0x60852a59 nnotes.DB2SearchFill@8+137

(2d7b;2d5f)

@[ 6] 0x60054cd4 nnotes.QueueFill@4+52 (2d7b)

@[ 7] 0x60054c11 nnotes.QueueGet@8+33

(2d7b;b0ef454)

@[ 8] 0x60066eee nnotes.GetNextBuffer@8+894

(3bfe0f8;60fb9d80)

@[ 9] 0x6004a85a nnotes.UpdateCollection@8+4282

(0;1)

@[10] 0x6007a587

nnotes.NIFUpdateCollectionNext@8+2439

(3c0a0f8;3c698f8)

@[11] 0x100418ce nserverl.ServerOpenCollection@8

+2286 (2a72164;b2780001)

@[12] 0x10021196 nserverl.DbServer@8+2486

(96bf0071;be040001)

@[13] 0x10035c8b nserverl.WorkThreadTask@8+1611

(646a8f4;0)

@[14] 0x10001a2e nserverl.Scheduler@4+750 (0)

@[15] 0x600fe6ef nnotes.ThreadWrapper@4+175 (0)

[16] 0x77e64829 kernel32.GetModuleHandleA+223

2

Subject: More info?

Sorry to hear that you are having problems, a few questions that may help get this solved:

  • What builddate&version of Domino is installed?

  • What OS type&version is this on?

  • What DB2 version&fixpack is installed?

  • Has the DB2 Access for Lotus Domino install also been upgraded?

  • Do you get an NSD report after the crash?

If it’s not something obvious (based on the answers to the above questions) you may want to open a PMR through support so this can be investigated further.

3

Subject: the obvious

Hi Pat, thanks for the tips and yea I covered those bases. All versions are as per Lotus requirements. DB2 is 9.1 fixpack 5 etc. I’m not an IBM business partner so cannot submit the NSD output and related files.

4

Subject: Take some NSDs during the CPU spike

If you see similar callstacks during the CPU spike (specifically the update collection calls), I have seen similar problems when the time stamp of a view or folder is set into the future.

5

Subject: pakistan

wierd subject? well it’s sort of strange cos it seems really well contained (not like the subject?)… i’ll move on. nserver.exe has now been consuming 25% of CPU cycles (4 CPU system) for over 2 days without degradation.

6

Subject: nserver 100%

I am having this same issue and have not found anything obvious. No problems until we upgraded to 8.5. Has anyone found an answer to this?

7

Subject: Same issue

We are having the same issue. I’ve been working with IBM since Friday (12/17) troubleshooting.

We are running 8.5.1.FP3

nserver.exe goes to 100% (or thereabouts) like clockwork every two hours on the odd hour (9am, 11am, 1pm, etc)

Killing all tasks has no effect

Restarting the server resolves the issue until the next odd hour

Even without amgr running, issue still occurs

Took server off-line and ran nfixup -F on entire data directory - issue still occurs

At a loss … any help would be appreciated.

Thanks,

  • Shane