This forum is closed to new posts and responses. Individual names altered for privacy purposes. The information contained in this website is provided for informational purposes only and should not be construed as a forum for customer support requests. Any customer support requests should be directed to the official HCL customer support channels below:

HCL Software Customer Support Portal for U.S. Federal Government clients
HCL Software Customer Support Portal


Mar 26, 2015, 3:00 PM
5 Posts

Transaction log effectivness -- is it performing well of not?

  • Category: Performance
  • Platform: Windows
  • Release: 9.0
  • Role: Administrator
  • Tags: Domino 9,Transaction Log
  • Replies: 5

We have a Domino system running custom business application (CRM, operations, billing, payments) with attachments with DAOS enabled.

130 active databases (with mail DBs), 5,5M documents

Daily we have mean 60 up to 110 concurent users,

Transaction Rate Mean Prime Shift=4 659, Up to 8000 peak

Documents Read Prime Shift=754 K

Documents Written Prime Shift=30 K 

We are wondering if our transaction log is performing well. We have 2Gb, Circular logging.

So we collected show stat database.rm.* statistics 4 times during the day to ask it it is performing well of not.

Here they are:

Statistic \\\\\ Daytime 09:25 11:25 13:29 17:27
Database.RM.Current.K.Restart.Redo.Hold 106650 122646 111232 68882
Database.RM.Current.K.UnCommited.Undo.Hold 0 0 0 3865
Database.RM.Current.MaxChkptScan.msec 620 620 620 620
Database.RM.Current.UnCommited.Time 03/26/2015 09:25:13 03/26/2015 11:26:33 03/26/2015 13:29:28 03/26/2015 17:25:01
Database.RM.LastChkpt.Avg.K.Logged.PerTran 16 18 30 1
Database.RM.LastChkpt.Interval.Sec 640 516 0 0
Database.RM.LastChkpt.K.Logged 41307 41305 91 26
Database.RM.LastChkpt.Time 03/26/2015 08:56:35 03/26/2015 11:26:11 03/26/2015 13:16:57 03/26/2015 17:25:24
Database.RM.LastChkpt.Total.DBs 402 402 120 247
Database.RM.LastChkpt.Trans 2568 2283 3 26
Database.RM.Logger.Buffers.Number.Waits 442966 443184 443254 498432
Database.RM.Logger.Buffers.Partial.Written 2 860 467 2 874 209 2 881 802 2 900 980
Database.RM.Logger.Buffers.Total.Written 19 664 575 19 808 026 19 896 331 20 105 147
Database.RM.Logger.Compress.Failure.LZ1.Bytes 163 021 132 163 144 012 163 476 108 163 897 788
Database.RM.Logger.Compress.Failure.LZ1.Number 8 483 8 489 8 511 8 534
Database.RM.Logger.Compress.Failure.Memory.Bytes 0 0 0 0
Database.RM.Logger.Compress.Failure.Memory.Number 0 0 0 0
Database.RM.Logger.Compress.Failure.Size.Bytes 31 810 141 31 869 333 31 960 971 32 064 881
Database.RM.Logger.Compress.Failure.Size.Number 37 182 37 329 37 512 37 794
Database.RM.Logger.Compress.RawBytes 32 929 547 305 33 168 180 748 33 465 385 364 33 868 274 345
Database.RM.Logger.Compress.Success.Bytes 6 971 915 716 7 019 910 779 7 085 747 962 7 176 488 948
Database.RM.Logger.Compress.Success.Number 1 606 485 1 618 183 1 632 714 1 652 464
Database.RM.Logger.IO.Avg.Write.Time 2 2 2 2
Database.RM.Logger.IO.Max.Write.Time 4371 4371 4371 4371
Database.RM.Logger.IO.Number.Writes 4 684 439 4 710 717 4 727 426 4 767 908
Database.RM.Media.Log.Recs.Read 0 0 0 0
Database.RM.Peak.Avg.K.Logged.PerTran 303 303 303 303
Database.RM.Peak.Interval.Sec 11 11 11 11
Database.RM.Peak.K.Logged 41310 41310 41310 41310
Database.RM.Peak.Time 03/19/2015 10:57:47 03/19/2015 10:57:47 03/19/2015 10:57:47 03/19/2015 10:57:47
Database.RM.Peak.Total.DBs 402 402 402 402
Database.RM.Peak.Trans 136 136 136 136
Database.RM.Peak.Trans.PerMin 741 741 741 741
Database.RM.Restart.Duration.Sec 17 17 17 17
Database.RM.Restart.K.Processed 144547 144547 144547 144547
Database.RM.Restart.Log.Recs.Analyzed 539589 539589 539589 539589
Database.RM.Restart.Log.Recs.Applied 1131 1131 1131 1131
Database.RM.Restart.Log.Recs.Read 772169 772169 772169 772169
Database.RM.Restart.Time 03/07/2015 01:06:03 03/07/2015 01:06:03 03/07/2015 01:06:03 03/07/2015 01:06:03
Database.RM.Restart.Total.DBs 5 5 5 5
Database.RM.Restart.Undo.Trans 0 0 0 0
Database.RM.SinceChkpt.K.Logged 24177 304 31818 361
Database.RM.SinceChkpt.Trans 2239 22 2020 647
Database.RM.SinceStartup.Aborts 1 1 1 1
Database.RM.SinceStartup.Critical.Log.Times 0 0 0 0
Database.RM.SinceStartup.Log.DiagRecs.Written 0 0 0 0
Database.RM.SinceStartup.Log.Recs.Applied 6 6 6 6
Database.RM.SinceStartup.Log.Recs.Read 10 10 10 10
Database.RM.SinceStartup.Log.Recs.Written 55313144 55551984 55729161 56198512
Database.RM.SinceStartup.M.Logged 65641 66147 66463 67203
Database.RM.SinceStartup.Trans 6028451 6055724 6087058 6162223
Database.RM.Sys.Log.Type Circular Circular Circular Circular
Database.RM.Sys.Logged Enabled Enabled Enabled Enabled
Database.RM.Sys.M.Chkpt.Interval 40 40 40 40
Database.RM.Sys.M.Log.Size 1984 1984 1984 1984
Database.RM.Sys.M.Redo.Limit 120 120 120 120
Database.RM.Sys.Phase Normal Normal Normal Normal
57 statistics found        
Mar 27, 2015, 8:57 AM
5 Posts
The question is following that topic

I read that technote and following it I cannot guess if my server has transaction log bottlenecks and can perform better, or bottlenecks in other parts of system. Technote doest give values to which one could compare his server.

I'l like to compare server statistics to other systems...

Exact questions:

would it be better to switch to Archive logging

of increasing log wil lead to longer time between commits

or whatever?

 

Can one help in this way?

Mar 27, 2015, 1:22 PM
32 Posts
The data looks fine but
this is scary:
Database.RM.Logger.IO.Max.Write.Time 4371 4371 4371 4371


4 seconds for a write ? It must have been a blip in the OS since the average is only 2. My only concern besides the max time is the size. I'd go w/ 4GB for the txn logs.

--Steve
Mar 27, 2015, 2:22 PM
5 Posts
Thanks Steve

We will increase Log to 4Gb and monitor Database.RM.Logger.IO.Max.Write.Time

Moreover, given server statistics are for 19 running days. So Database.RM.SinceStartup.M.Logged divided by 19 gives 3,5Gb daily

So we will have Log volume enough for a whole day.

Apr 1, 2015, 7:02 AM
33 Posts
Question

Is your transaction log on a separate harddisk.

If not, that is a big bottle neck right there.

Serial write/ - Random Read/Write switching kills disk performance.


This forum is closed to new posts and responses. Individual names altered for privacy purposes. The information contained in this website is provided for informational purposes only and should not be construed as a forum for customer support requests. Any customer support requests should be directed to the official HCL customer support channels below:

HCL Software Customer Support Portal for U.S. Federal Government clients
HCL Software Customer Support Portal