RepServer Throughput Script - Revision history

Bobh at 11:29, 26 February 2009

2009-02-26T11:29:41Z

Bobh at 11:25, 26 February 2009

2009-02-26T11:25:55Z

Bobh at 11:22, 26 February 2009

2009-02-26T11:22:17Z

Bobh at 11:15, 20 February 2009

2009-02-20T11:15:17Z

Bobh at 11:14, 20 February 2009

2009-02-20T11:14:46Z

Bobh at 09:44, 20 February 2009

2009-02-20T09:44:48Z

Bobh at 09:41, 20 February 2009

2009-02-20T09:41:31Z

Bobh at 00:27, 20 February 2009

2009-02-20T00:27:49Z

Bobh at 00:26, 20 February 2009

2009-02-20T00:26:38Z

Bobh at 00:05, 20 February 2009

2009-02-20T00:05:28Z

@@ Line 141: / Line 141: @@
 #date
 #echo "Queue Name, First segment, Last segment, Next segment"
-cat $SQMFILE | egrep $insert_filter | sed 's/Awaiting Message/AwaitingMessage/; s/Awaiting Wakeup/AwaitingWakeup/; s/\(\.[0-9][0-
+cat $SQMFILE | egrep $insert_filter | sed 's/Awaiting Message/AwaitingMessage/; s/Awaiting Wakeup/AwaitingWakeup/; s/\(\.[0-9][0-9]*\)\.[0-9]/\1/' | grep -v ":0 ${primary_ds}" | awk '{print $4,$14,$15,$16}' | while read q_name first_seg last_seg next_seg
-−
 ]*\)\.[0-9]/\1/' | grep -v ":0 ${primary_ds}" | awk '{print $4,$14,$15,$16}' | while read q_name first_seg last_seg next_seg
 do
    q_size=$(bcalc "$last_seg - $first_seg")
@@ Line 202: / Line 200: @@
 declare @eta datetime
 select convert(char(19),sample_time) sample_time, q_name, size q_size, mb_processed, thruput "thruput (mb/min)"  from rs_throughput
 where sample_time >= dateadd(mi,-${option2},getdate()) and q_name like "%${q_name}%"
@@ Line 272: / Line 267: @@
 eof
 printf "Current Stable Queue usage:\n"
-cat $SQMFILE | egrep $insert_filter | sed 's/Awaiting Message/AwaitingMessage/; s/Awaiting Wakeup/AwaitingWakeup/; s/\(\.[0-9][0-
+cat $SQMFILE | egrep $insert_filter | sed 's/Awaiting Message/AwaitingMessage/; s/Awaiting Wakeup/AwaitingWakeup/; s/\(\.[0-9][0-9]*\)\.[0-9]/\1/' | grep -v ":0 ${primary_ds}" | awk '{print $4,$14,$15,$16}' | while read q_name first_seg last_seg next_seg
-−
 ]*\)\.[0-9]/\1/' | grep -v ":0 ${primary_ds}" | awk '{print $4,$14,$15,$16}' | while read q_name first_seg last_seg next_seg
 do
    q_size=$(bcalc "$last_seg - $first_seg")

@@ Line 82: / Line 82: @@
                                        # !! for your own environment.
 RISQLCMD="isql -Usa -P${RSPASS} -S${REP_SRV} -w1000"
-ISQLCMD="isql -Usa -P${SYBPASS} -S${RPT_SRV} -w200 -D${RPT_DB}"
+ISQLCMD="isql -Usa -P${ASEPASS} -S${RPT_SRV} -w200 -D${RPT_DB}"

@@ Line 72: / Line 72: @@
 RECIPIENTS="user@domain"               # !! email config (separate addresses with commas)
 ADMINDIR="/opt/home/sybase/admin"      # !! sybase admin directory
 # static:
@@ Line 79: / Line 81: @@
 # config aware environment variables: # !! You may need to replace the line below with the equivalent
                                        # !! for your own environment.
-RISQLCMD="isql -Usa -P`grep "$REP_SRV," $HOME/admin/.servers | cut  -d , -f 4` -S$REP_SRV -w1000"
+RISQLCMD="isql -Usa -P${RSPASS} -S${REP_SRV} -w1000"
 ISQLCMD="isql -Usa -P${SYBPASS} -S${RPT_SRV} -w200 -D${RPT_DB}"

@@ Line 12: / Line 12: @@
 ) You will need to cron this script with the insert option (rs_throughput.sh insert) and allow it to collect at least 2 samples before being able to run it interactively to obtain throughput information.  The script runs silently when run with the insert option.<br>
 ) The script automatically determines the elapsed interval in minutes between samples so you can cron the script to your preferred interval, be it 3 minutes or 10 minutes, the mb/min figure will be calculated correctly.<br>
-) In addition to configuring the script to your environment, you will also need to create the rs_throughput table prior to first run.
+) In addition to configuring the script to your environment, you will also need to create the rs_throughput table prior to first run.<br>
 ) Re-sync ETA: this value is provided based on the current throughput average for the period specified.  For example, this means that running rs_throughput.sh 15 may give a different Re-Sync ETA compared to running rs_throughput.sh 20
 <pre>

@@ Line 3: / Line 3: @@
 o average throughput per queue (mb per minute)<br>
 o average throughput combined (all queues)<br>
-o estimated time for a replicate to catch up<br>
+o estimated time for a replicate to catch up (Re-sync ETA)<br>
 o current queue sizes<br>
 There is also a maximum recorded throughput option but this appears not to work well as some of the figures have been suspicious.<br>
@@ Line 13: / Line 13: @@
 ) The script automatically determines the elapsed interval in minutes between samples so you can cron the script to your preferred interval, be it 3 minutes or 10 minutes, the mb/min figure will be calculated correctly.<br>
 ) In addition to configuring the script to your environment, you will also need to create the rs_throughput table prior to first run.
 <pre>
 #!/usr/bin/ksh

@@ Line 8: / Line 8: @@
 The real maximum throughput can be established by noting when the replicates start to fall behind in conjunction with the current throughput figures.<br>
 This script can be particularly useful in helping to ascertain if you need to throw more CPU at your repserver box.
 <pre>
 #!/usr/bin/ksh

@@ Line 23: / Line 23: @@
 # Modification history:
 # 19/02/09 : Bob Holmes : Script modified to enable easier customisation prior
-#                         to release on Sybase wiki. (modification not tested)
+#                         to release on Sybase wiki. (modifications not tested)
 # --------------------------
 # Comments: This script relies on the reporting..rs_throughput table for data