Notes hates one user-update

One user’s Notes experience is really really really slow compared to everyone else.

Her network settings are not the issue. If we move her to another PC, she sees the same slowness. If I switch to her ID on my computer, I see the same slowness. If I switch to another ID that has the same rights as hers, the slowness goes away.

I’ve recreated her ID thinking hers might be very old and somehow looking for an old server we no longer use - no joy.

I’m having IT re-install her Notes client from scratch but don’t expect this to help given that I see the same problem on my PC with her newly created ID.

This is in the Notes client, single location, one production server (we have a cluster server but it is set to deny access to users unless the main server goes down).

There is nothing unusual about her Notes, network, or PC setup compared to others (and none of that matters since we see the problem wherever her ID is used).

There is nothing in the apps keyed to her name - they all use ACL roles and she’s in very common group that many folks belong to.

The problem started a few weeks back and we can’t find anything that seems to correlate to the change in performance.

I don’t know what else to look at. Anybody have any ideas?

TIA

Doug

I thought I’d add an example:

I have a fairly old design that pre-dates Native document locking. When a doc is opened, a lock document is created that contains the doc ID and the user name. If anybody tries to edit the doc, the lock is found and the edit denied. When the lock owner exits the doc, the QueryClose event kills the lock. For a normal user, the doc close/kill lock process takes less than a second. For this user, it runs up to 15 seconds.

QueryClose code:

REM { Unlocks document on exit};

LockKey := @Text(@DocumentUniqueID);

IsLockedBy := @DbLookup(“Notes”:“NoCache”;“”;“(LockDocs)”;LockKey;2);

@If(@IsError(IsLockedBy);@Return(“”);@Success);

Server := @Subset(@Name([CN]; @DbName); 1) ;

DB := @Subset(@DbName; -1);

@If(

IsLockedBy != @Name([CN];@UserName);

@Success;

@Do(

@Command([FileOpenDatabase];Server:DB;“(LockDocs)”; Lockkey ; “1” ; “1” );

@Command([ToolsRunMacro];“(Hide)”);

@Command([FileCloseWindow])))

Hide Macro code:

FIELD Hide := Hide;

FIELD LockOff := LockOff;

@SetField(“Hide”;“1”);

@SetField(“LockOff”;@Now);

SELECT @All

Nothing magic there and the slowness only affects this one user.

I’ve been able to narrow it down a bit - she’s dead slow when opening or changing views. The code that creates the lock doc doesn’t reference any views and runs as fast as anything else. The QueryClose code opens views and is slow. If she just changes views, it’s slow - 5 - 10x slower than a typical user. It’s like she’s searching the web for the right view…

Subject: Resolved - corrupt db

I ran fixup on the db last night after everyone was off-line. Poof - magic. My problem id is no longer a problem id.

I had run fixup during the day via the admin client, but I think it did not run since users were connected to the db. Repeating the fixup routine with nobody running the db seemed to do the trick.

Gee, that was fun!

My thanks to everyone who posted suggestions.

Doug

Subject: RE: Resolved - corrupt db

Great news, thank for the follow up!

JYR

Subject: Notes loves her too much is more like it

Doug,Single user issues are always hard to resolve, but try these options:

First check for some agents she has running on her mail file that run at startup or her private views may be rebuilding on the fly.

You can check the latter in your domino log.

Run FUC on her mail file.(Fixup, Updall,Compact)

Lastly how biog is her mail file or more importantly how big is her inbox? These can also cause delays.

Delete her full text index on her mail file and recreate it as well.

Turn off auto index if it is running on her mail file.

Users are their worst enemies once they try to do something without proper advice or reading the manual.

You can email me keith at vanessabrooks.com or IM me via bleedyellow .com

Subject: Swing and a miss…

No mail file to check. We switched to Outlook/Exchange about 3 years back and her account was created after the move.

I did another check - the problem does seem to be localized to a single db. If I have her ID open another db based on the same template, the performance is the same as other users…views pop open, the doc open/close code executes at light speed.

Maybe the db is broken…may be time for fixup/updall…I’ll let you know what I find.

Subject: RE: Swing and a miss…

Have you deleted her cache.ndk?

Subject: RE: Swing and a miss…

cachebookmark

desktop6

cleaned the ini file down to 3 lines and did what I’d call a re-setup.

If I switch to her ID on my PC (so I’m totally disconnected from her typical working environment) I see the same issue.

Check my response to JYR for more debug data.

Thanks for the reply. Keep 'em coming…

Doug

Subject: RE: Swing and a miss…

Any readers fields?

JYR

Subject: No reader fields

Subject: Client Performance Checklist

Hi Doug,

From Notes/Domino Best Practices: Performance

http://www-1.ibm.com/support/docview.wss?rs=463&uid=swg27008849

                   Client Performance Checklist 

JYR

Subject: RE: Client Performance Checklist

Thanks for the links. I’ve done a quick check on the client performance issues and don’t see anything obvious. All of the items either refer to the application design or to a generic client setup.

In this case, the application performance is user dependant but I can’t find anything tied to this particular user.

The client setup isn’t at issue since the problem follows the user’s ID file from pc to pc; if it was setup, then my using her id on my pc wouldn’t show a performance problem, but it does.

I’ll spend some more time poking through the references.

Thanks for sticking with this; it’s very much appreciated.

Doug

Subject: RE: Client Performance Checklist

HI,

Format c:?

Remove, compact the workspace and readd icon, i know it funny but just in case.

Is there any special feature (Select in views, etc) in this db?

Any private views?

Any special access for this user?

Lookup on other db that the user might not have full access?

Any profile documents in your db?

If you run the Client_clock, do you see something special?

Any error messages?

JYR

Subject: RE: Client Performance Checklist

I like the format option… :wink: But, darn, it has nothing to do with her drive - the issue follows her ID from pc to pc so formatting her drive won’t do anything since I can replicate the issue on my pc by switching to her ID file.

Is there any special feature (Select in views, etc) in this db?

Nothing but standard select form = x type stuff

Any private views?

Nope. I don’t even let users create private views.

Any special access for this user?

Nope. She’s in a single group ‘A-L’ based on her last name just like everybody else who’s name starts with A to L. Editor access to the db just like everyone else. No reader fields. No funky view code looking for her, nothing.

Lookup on other db that the user might not have full access?

Nothing. This wouldn’t be the issue since views can’t do lookups and the performance hit we’re seeing is when she opens views in this one db.

Any profile documents in your db?

None.

If you run the Client_clock, do you see something special?

Any error messages?

I haven’t tried this - I’ll see if I can do it this afternoon. When I looked at the documentation on the Client_Clock, it didn’t seem like it would tell me much. I know the view loads are long and this will just tell me how long. It won’t tell me why the loads are long so I’m not clear on what I’ll gain from running the Clock. If I were trying to tune the application, it would be great since I could see where I was paying for a feature but in this case, the view load times are dependent not on the design, but the user’s name in ONE db based on a template. Switch dbs (same design, different content), the problem goes away…???

Subject: Speed test-updated

Workflow:Login as problem user on my pc

Open db to View 1

Switch to View 2

Switch to View 1

Exit db

Exit Notes

Change to my id and repeat.

Change to random other user with same group membership as problem user and repeat test

Chart below only has the first two tests.

Problem user

SpeedTrap_D-9K9GR61_2008_03_06@13_18_51.txt

(1-8 [1]) OPEN_DB(CN=Noteskmc/O=Kollsman!!Kmc2\wizzer\wipWizzer.nsf): (Connect to Noteskmc/Kollsman: 0 ms) (Exch names: 0 ms)(Authenticate: 515 ms.)

(OPEN_SESSION: 16 ms)

31 ms. [134+290=424]

(2-9 [2]) GET_UNREAD_NOTE_TABLE: 16 ms. [290+85196=85486]

(3-9 [3]) OPEN_NOTE(REP852572E2:00694556-NTFFFF0010,03000400): 0 ms. [48+11660=11708]

(4-9 [4]) OPEN_COLLECTION(REP852572E2:00694556-NT000002B2,0040,0000): RCV_UNREAD[85070] 5281 ms. [250+85204=85454]

(5-14 [5]) GET_NOTE_INFO: 0 ms. [18+102=120]

(6-14 [6]) SET_COLLATION: 0 ms. [12+14=26]

(7-14 [7]) UPDATE_FILTERS(REP852572E2:00694556-NT000002B2): 0 ms. [42+12=54]

(8-14 [8]) READ_ENTRIES(REP852572E2:00694556-NT000002B2): 16 ms. [76+9558=9634]

(9-14 [9]) OPEN_NOTE(REP852572E2:00694556-NTFFFF0040,00000400): 16 ms. [48+1532=1580]

(10-22 [10]) OPEN_COLLECTION(REP852572E2:00694556-NT000002BA,0040,0000): RCV_UNREAD[85070] 5140 ms. [250+85204=85454]

(11-27 [11]) CLOSE_COLLECTION(REP852572E2:00694556-NT000002B2): 0 ms. [12+0=12]

(12-27 [12]) GET_NOTE_INFO: 0 ms. [18+296=314]

(13-27 [13]) SET_COLLATION: 0 ms. [12+14=26]

(14-27 [14]) READ_ENTRIES(REP852572E2:00694556-NT000002BA): 16 ms. [88+10466=10554]

(15-27 [15]) READ_ENTRIES(REP852572E2:00694556-NT000002BA): 15 ms. [92+11558=11650]

(16-30 [16]) OPEN_COLLECTION(REP852572E2:00694556-NT000002B2,0040,0000): RCV_UNREAD[85070] 5141 ms. [250+85204=85454]

(17-35 [17]) CLOSE_COLLECTION(REP852572E2:00694556-NT000002BA): 0 ms. [12+0=12]

(18-35 [18]) GET_NOTE_INFO: 188 ms. [18+102=120]

(19-35 [19]) SET_COLLATION: 0 ms. [12+14=26]

(20-35 [20]) UPDATE_FILTERS(REP852572E2:00694556-NT000002B2): 0 ms. [42+12=54]

(21-35 [21]) READ_ENTRIES(REP852572E2:00694556-NT000002B2): 0 ms. [76+9242=9318]

(22-35 [22]) READ_ENTRIES(REP852572E2:00694556-NT000002B2): 16 ms. [76+9582=9658]

(23-37 [23]) CLOSE_COLLECTION(REP852572E2:00694556-NT000002B2): 0 ms. [12+0=12]

(24-37 [24]) CLOSE_DB(REP852572E2:00694556): 0 ms. [14+0=14]

0 ms. [1623863594+41710256=1665573850] (Session Closed)

Me

SpeedTrap_D-9K9GR61_2008_03_06@13_20_50.txt

(1-5 [1]) OPEN_DB(CN=Noteskmc/O=Kollsman!!Kmc2\wizzer\wipWizzer.nsf): (Connect to Noteskmc/Kollsman: 0 ms) (Exch names: 0 ms)(Authenticate: 0 ms.)

(OPEN_SESSION: 0 ms)

15 ms. [134+290=424]

(2-6 [2]) GET_UNREAD_NOTE_TABLE: 0 ms. [290+34876=35166]

(3-6 [3]) OPEN_NOTE(REP852572E2:00694556-NTFFFF0010,03000400): 0 ms. [48+11908=11956]

(4-6 [4]) OPEN_COLLECTION(REP852572E2:00694556-NT000002B2,0040,0000): RCV_UNREAD[34810] 31 ms. [250+34884=35134]

(5-6 [5]) GET_NOTE_INFO: 0 ms. [18+102=120]

(6-6 [6]) SET_COLLATION: 0 ms. [12+14=26]

(7-6 [7]) UPDATE_FILTERS(REP852572E2:00694556-NT000002B2): 0 ms. [42+12=54]

(8-6 [8]) READ_ENTRIES(REP852572E2:00694556-NT000002B2): 16 ms. [76+9558=9634]

(9-6 [9]) OPEN_NOTE(REP852572E2:00694556-NTFFFF0040,00000400): 0 ms. [48+1780=1828]

(10-12 [10]) OPEN_COLLECTION(REP852572E2:00694556-NT000002BA,0040,0000): RCV_UNREAD[34810] 78 ms. [250+34884=35134]

(11-12 [11]) CLOSE_COLLECTION(REP852572E2:00694556-NT000002B2): 0 ms. [12+0=12]

(12-12 [12]) GET_NOTE_INFO: 93 ms. [18+102=120]

(13-12 [13]) SET_COLLATION: 0 ms. [12+14=26]

(14-12 [14]) READ_ENTRIES(REP852572E2:00694556-NT000002BA): 0 ms. [88+10466=10554]

(15-12 [15]) READ_ENTRIES(REP852572E2:00694556-NT000002BA): 16 ms. [92+11558=11650]

(16-18 [16]) OPEN_COLLECTION(REP852572E2:00694556-NT000002B2,0040,0000): RCV_UNREAD[34810] 31 ms. [250+34884=35134]

(17-18 [17]) CLOSE_COLLECTION(REP852572E2:00694556-NT000002BA): 0 ms. [12+0=12]

(18-18 [18]) GET_NOTE_INFO: 125 ms. [18+102=120]

(19-18 [19]) SET_COLLATION: 0 ms. [12+14=26]

(20-18 [20]) UPDATE_FILTERS(REP852572E2:00694556-NT000002B2): 0 ms. [42+12=54]

(21-18 [21]) READ_ENTRIES(REP852572E2:00694556-NT000002B2): 0 ms. [76+9242=9318]

(22-18 [22]) READ_ENTRIES(REP852572E2:00694556-NT000002B2): 16 ms. [76+9582=9658]

(23-20 [23]) CLOSE_COLLECTION(REP852572E2:00694556-NT000002B2): 0 ms. [12+0=12]

(24-21 [24]) CLOSE_DB(REP852572E2:00694556): 0 ms. [14+0=14]

0 ms. [1623864208+41918810=1665783018] (Session Closed)

Interpretation:

My problem user takes just over 5 seconds to open views.

I take 0.05 seconds on average to open the same views.

Random other user in the same group as my problem user (not shown in this test set) were at about 78 ms so in line with my rates.

Problem user has a problem…

Subject: RE: Speed test

Seems to have a lot of difference with unread documents

The format of an RPC entry in the log is (Sequential Number), RPC_NAME, time to complete, [bytes_sent+bytes_received=total].

User

SpeedTrap_D-9K9GR61_2008_03_06@13_18_51.txt
(1-8 [1]) OPEN_DB(CN=Noteskmc/O=Kollsman!!Kmc2\wizzer\wipWizzer.nsf): (Connect to Noteskmc/Kollsman: 0 ms) (Exch names: 0 ms)(Authenticate: 515 ms.)
(OPEN_SESSION: 16 ms)
31 ms. [134+290=424]
(2-9 [2]) GET_UNREAD_NOTE_TABLE: 16 ms. [290+85196=85486]
(3-9 [3]) OPEN_NOTE(REP852572E2:00694556-NTFFFF0010,03000400): 0 ms. [48+11660=11708]
(4-9 [4]) OPEN_COLLECTION(REP852572E2:00694556-NT000002B2,0040,0000): RCV_UNREAD[85070] 5281 ms. [250+85204=85454]
(5-14 [5]) GET_NOTE_INFO: 0 ms. [18+102=120]
(6-14 [6]) SET_COLLATION: 0 ms. [12+14=26]
(7-14 [7]) UPDATE_FILTERS(REP852572E2:00694556-NT000002B2): 0 ms. [42+12=54]
(8-14 [8]) READ_ENTRIES(REP852572E2:00694556-NT000002B2): 16 ms. [76+9558=9634]
(9-14 [9]) OPEN_NOTE(REP852572E2:00694556-NTFFFF0040,00000400): 16 ms. [48+1532=1580]
(10-22 [10]) OPEN_COLLECTION(REP852572E2:00694556-NT000002BA,0040,0000): RCV_UNREAD[85070] 5140 ms. [250+85204=85454]
(11-27 [11]) CLOSE_COLLECTION(REP852572E2:00694556-NT000002B2): 0 ms. [12+0=12]
(12-27 [12]) GET_NOTE_INFO: 0 ms. [18+296=314]
(13-27 [13]) SET_COLLATION: 0 ms. [12+14=26]
(14-27 [14]) READ_ENTRIES(REP852572E2:00694556-NT000002BA): 16 ms. [88+10466=10554]
(15-27 [15]) READ_ENTRIES(REP852572E2:00694556-NT000002BA): 15 ms. [92+11558=11650]
(16-30 [16]) OPEN_COLLECTION(REP852572E2:00694556-NT000002B2,0040,0000): RCV_UNREAD[85070] 5141 ms. [250+85204=85454]
(17-35 [17]) CLOSE_COLLECTION(REP852572E2:00694556-NT000002BA): 0 ms. [12+0=12]
(18-35 [18]) GET_NOTE_INFO: 188 ms. [18+102=120]
(19-35 [19]) SET_COLLATION: 0 ms. [12+14=26]
(20-35 [20]) UPDATE_FILTERS(REP852572E2:00694556-NT000002B2): 0 ms. [42+12=54]
(21-35 [21]) READ_ENTRIES(REP852572E2:00694556-NT000002B2): 0 ms. [76+9242=9318]
(22-35 [22]) READ_ENTRIES(REP852572E2:00694556-NT000002B2): 16 ms. [76+9582=9658]
(23-37 [23]) CLOSE_COLLECTION(REP852572E2:00694556-NT000002B2): 0 ms. [12+0=12]
(24-37 [24]) CLOSE_DB(REP852572E2:00694556): 0 ms. [14+0=14]
0 ms. [1623863594+41710256=1665573850] (Session Closed)

Me
SpeedTrap_D-9K9GR61_2008_03_06@13_20_50.txt
(1-5 [1]) OPEN_DB(CN=Noteskmc/O=Kollsman!!Kmc2\wizzer\wipWizzer.nsf): (Connect to Noteskmc/Kollsman: 0 ms) (Exch names: 0 ms)(Authenticate: 0 ms.)
(OPEN_SESSION: 0 ms)
15 ms. [134+290=424]
(2-6 [2]) GET_UNREAD_NOTE_TABLE: 0 ms. [290+34876=35166]
(3-6 [3]) OPEN_NOTE(REP852572E2:00694556-NTFFFF0010,03000400): 0 ms. [48+11908=11956]
(4-6 [4]) OPEN_COLLECTION(REP852572E2:00694556-NT000002B2,0040,0000): RCV_UNREAD[34810] 31 ms. [250+34884=35134]
(5-6 [5]) GET_NOTE_INFO: 0 ms. [18+102=120]
(6-6 [6]) SET_COLLATION: 0 ms. [12+14=26]
(7-6 [7]) UPDATE_FILTERS(REP852572E2:00694556-NT000002B2): 0 ms. [42+12=54]
(8-6 [8]) READ_ENTRIES(REP852572E2:00694556-NT000002B2): 16 ms. [76+9558=9634]
(9-6 [9]) OPEN_NOTE(REP852572E2:00694556-NTFFFF0040,00000400): 0 ms. [48+1780=1828]
(10-12 [10]) OPEN_COLLECTION(REP852572E2:00694556-NT000002BA,0040,0000): RCV_UNREAD[34810] 78 ms. [250+34884=35134]
(11-12 [11]) CLOSE_COLLECTION(REP852572E2:00694556-NT000002B2): 0 ms. [12+0=12]
(12-12 [12]) GET_NOTE_INFO: 93 ms. [18+102=120]
(13-12 [13]) SET_COLLATION: 0 ms. [12+14=26]
(14-12 [14]) READ_ENTRIES(REP852572E2:00694556-NT000002BA): 0 ms. [88+10466=10554]
(15-12 [15]) READ_ENTRIES(REP852572E2:00694556-NT000002BA): 16 ms. [92+11558=11650]
(16-18 [16]) OPEN_COLLECTION(REP852572E2:00694556-NT000002B2,0040,0000): RCV_UNREAD[34810] 31 ms. [250+34884=35134]
(17-18 [17]) CLOSE_COLLECTION(REP852572E2:00694556-NT000002BA): 0 ms. [12+0=12]
(18-18 [18]) GET_NOTE_INFO: 125 ms. [18+102=120]
(19-18 [19]) SET_COLLATION: 0 ms. [12+14=26]
(20-18 [20]) UPDATE_FILTERS(REP852572E2:00694556-NT000002B2): 0 ms. [42+12=54]
(21-18 [21]) READ_ENTRIES(REP852572E2:00694556-NT000002B2): 0 ms. [76+9242=9318]
(22-18 [22]) READ_ENTRIES(REP852572E2:00694556-NT000002B2): 16 ms. [76+9582=9658]
(23-20 [23]) CLOSE_COLLECTION(REP852572E2:00694556-NT000002B2): 0 ms. [12+0=12]
(24-21 [24]) CLOSE_DB(REP852572E2:00694556): 0 ms. [14+0=14

Subject: RE: Speed test

Sorry for being dense but ‘Seems to have a lot of difference with unread documents’ means what in relation to this problem?

If I look at the Db properties for a ‘fast’ and ‘slow’ db for this user, the settings for unread marks are the same (the ‘don’t keep unread marks’ option is not selected on both dbs).

Since I’m running her id on my client, there are no person-specific preferences that would come into play, right?

Does the speed difference point to something I should look at?

FWIW, the dbs don’t use unread marks.

Doug

Subject: Unread marks-corrupt db - update

When I look at the db icon on my desktop, I see about 33K unread docs on the icon.When I switch to the problem user’s ID, the unread count goes to xxxx

Total docs in the db are about 33K

She’s in the db much more than me so you’d think the unread counts for her would be the same or less than for me.

What would cause the unread indicator to be all x’s?

When I open a different db with her ID, the unread counts on the icon remain visible (about 9.9K) and match the numbers I see with my id.

I’m making a copy of the db to see if maybe the db is corrupted. I’ll let you know what I find.

Update - the copy works fine for the user. It looks like there is something busted inside the current db.

Subject: RE: Unread marks-corrupt db - update

Hi Doug,

Normally, XXXXX means more than 100 000 docs.

Can you create a view Select @all

first colum - Value: Form Sort : Categorize

second column , get the total of document of the first column

does the total is identical for you and for her?

UPDATE: mmmmm, very weird.!! :slight_smile:

JYR

Subject: Doc counts

Doc counts match one for one.

I created private views under my ID and hers but both show 32583 docs.

Select @All, don’t show response in heirarchy. Column 1 = the number 1 totaled don’t show sub totals, Column 2 = form, sorted categorized.

I feel like I’m living in flatland, and I’m just a poor little dot that can’t conceive the 3rd dimension…

Subject: RE: Doc counts

I feel like I’m living in flatland.