This post is more than 5 years old
2 Intern
•
1.3K Posts
0
1436
April 16th, 2009 18:00
What are the advantage of maintaining the Power Path
AFAIK, the mainteance on Power Path is only the "version upgrade" periodically to keep up with the EOSL or to fight with the effected bugs. PP upgrade(requires reboot) we normally propose with other system related acitivities to avoid a separate outages. But I have noticed that not much people really cares about the EOSL or the bugs and we are still have systems with PP versions 3.0.3 on critical production servers.
Do micro code upgrade of SYM or flare code upgrade on CLAR recommends to keep the Sulution Enabler and PP (which ever applicable) updated. I am thinking why the PP left unupgraded even though the codes on array are updated.
Can i say PP is an integral part of the SAN ?Maintaining PP and array should get imprtance and should be under same umberla; is it??
I am concerned about PP as we have seen many possible bugs pointing to PP with old versions which effect systems availability. Here is one case(with version 4.5) where multiple threads that are hung in a set of EMC Powerpath code routines while trying to do I/Os.
Here is how stack trace looks like
proc_list{47} pid=77 tid=131 cmd="MpcPeriodicCallbackDaemon"
p_stat="SINUSE" : kt_stat="TSSLEEP"
Process : p proc_t 0x103d6a040
pid=77 MpcPeriodicCal
Kthread : p kthread_t 0x103d6b040
Using PCB: p user_t 0xff37400.0x400003ffffff0000
SR5=0xff37400
LVL FUNC ARG0 ARG1 ARG2 ARG3 ARG4 ARG5 ARG6 ARG7
0) _sleep+0x380 0x103d39390 0x294 n/a n/a n/a n/a n/a n/a
1) PowerSleep+0x6c 0x103d39390 0x103d39358 n/a n/a n/a n/a n/a n/a
2) PowerGetSema+0x1c0 0x103d39358 n/a n/a n/a n/a n/a n/a n/a
3) PowerServiceDaemonQ+0x34 0x103d39300 n/a n/a n/a n/a n/a n/a n/a
4) PowerPlatformCreateDaemonProc+0xac n/a n/a n/a n/a n/a n/a n/a n/a
5) PowerServiceDaemonQ+0xe8 0xebb10400 n/a n/a n/a n/a n/a n/a n/a
6) PowerPlatformCreateDaemonProc+0xac n/a n/a n/a n/a n/a n/a n/a n/a
7) PiocDaemon+0x118 0x40004000 0x262a5837 n/a n/a n/a n/a n/a n/a
8) power_pseudo_ioctl+0xcc n/a 0xffffffff80105060 0x103cea040 0x3 n/a n/a n/a n/a
9) spec_ioctl+0xac 0x103cab048 0xffffffff80105060 0x103cea040 0x3 n/a n/a n/a n/a
10) vno_ioctl+0x98 0x86a2faf8 0xffffffff80105060 0x103cea040 n/a n/a n/a n/a 0xb8900d0
11) ioctl+0x118 0x400003ffffff03a0 n/a n/a n/a n/a n/a n/a 0x8a2464c
12) syscall+0x204 n/a n/a n/a n/a n/a n/a n/a n/a
13) syscallinit+0x55c n/a n/a n/a n/a n/a n/a n/a n/a
Other way of saying alot of Powerpath related I/Os that seem to be stuck
LVL FUNC
0) _sleep+0x380
1) PowerSleep+0x6c
2) PowerGetSema+0x1c0
3) PowerServiceDaemonQ+0x34
4) PowerPlatformCreateDaemonProc+0xac
5) PowerServiceDaemonQ+0xe8
6) PowerPlatformCreateDaemonProc+0xac
7) PiocDaemon+0x118
8) power_pseudo_ioctl+0xcc
9) spec_ioctl+0xac
10) vno_ioctl+0x98
11) ioctl+0x118
12) syscall+0x204
13) syscallinit+0x55c
I am curious who make sure(owns) that PP versions are updated on your shops???Why or why NOT maintaning PP should be an integral part of keeping a SAN healthy?
Do micro code upgrade of SYM or flare code upgrade on CLAR recommends to keep the Sulution Enabler and PP (which ever applicable) updated. I am thinking why the PP left unupgraded even though the codes on array are updated.
Can i say PP is an integral part of the SAN ?Maintaining PP and array should get imprtance and should be under same umberla; is it??
I am concerned about PP as we have seen many possible bugs pointing to PP with old versions which effect systems availability. Here is one case(with version 4.5) where multiple threads that are hung in a set of EMC Powerpath code routines while trying to do I/Os.
Here is how stack trace looks like
proc_list{47} pid=77 tid=131 cmd="MpcPeriodicCallbackDaemon"
p_stat="SINUSE" : kt_stat="TSSLEEP"
Process : p proc_t 0x103d6a040
pid=77 MpcPeriodicCal
Kthread : p kthread_t 0x103d6b040
Using PCB: p user_t 0xff37400.0x400003ffffff0000
SR5=0xff37400
LVL FUNC ARG0 ARG1 ARG2 ARG3 ARG4 ARG5 ARG6 ARG7
0) _sleep+0x380 0x103d39390 0x294 n/a n/a n/a n/a n/a n/a
1) PowerSleep+0x6c 0x103d39390 0x103d39358 n/a n/a n/a n/a n/a n/a
2) PowerGetSema+0x1c0 0x103d39358 n/a n/a n/a n/a n/a n/a n/a
3) PowerServiceDaemonQ+0x34 0x103d39300 n/a n/a n/a n/a n/a n/a n/a
4) PowerPlatformCreateDaemonProc+0xac n/a n/a n/a n/a n/a n/a n/a n/a
5) PowerServiceDaemonQ+0xe8 0xebb10400 n/a n/a n/a n/a n/a n/a n/a
6) PowerPlatformCreateDaemonProc+0xac n/a n/a n/a n/a n/a n/a n/a n/a
7) PiocDaemon+0x118 0x40004000 0x262a5837 n/a n/a n/a n/a n/a n/a
8) power_pseudo_ioctl+0xcc n/a 0xffffffff80105060 0x103cea040 0x3 n/a n/a n/a n/a
9) spec_ioctl+0xac 0x103cab048 0xffffffff80105060 0x103cea040 0x3 n/a n/a n/a n/a
10) vno_ioctl+0x98 0x86a2faf8 0xffffffff80105060 0x103cea040 n/a n/a n/a n/a 0xb8900d0
11) ioctl+0x118 0x400003ffffff03a0 n/a n/a n/a n/a n/a n/a 0x8a2464c
12) syscall+0x204 n/a n/a n/a n/a n/a n/a n/a n/a
13) syscallinit+0x55c n/a n/a n/a n/a n/a n/a n/a n/a
Other way of saying alot of Powerpath related I/Os that seem to be stuck
LVL FUNC
0) _sleep+0x380
1) PowerSleep+0x6c
2) PowerGetSema+0x1c0
3) PowerServiceDaemonQ+0x34
4) PowerPlatformCreateDaemonProc+0xac
5) PowerServiceDaemonQ+0xe8
6) PowerPlatformCreateDaemonProc+0xac
7) PiocDaemon+0x118
8) power_pseudo_ioctl+0xcc
9) spec_ioctl+0xac
10) vno_ioctl+0x98
11) ioctl+0x118
12) syscall+0x204
13) syscallinit+0x55c
I am curious who make sure(owns) that PP versions are updated on your shops???Why or why NOT maintaning PP should be an integral part of keeping a SAN healthy?
No Events found!
AranH1
2.2K Posts
0
April 17th, 2009 08:00
As far as how up to date I keep the installations, well it all depends on the versions. It seems like there is one really stable and recommended version within every major release (4.5.1 and 5.1.0.51 seem to be really solid). So I tend to keep all the hosts on the most stable revision within one or two major revisions. I never go to the latest revision as there always seems to be issues that I let other people figure out before putting it on my hosts
RRR
2 Intern
•
5.7K Posts
1
April 17th, 2009 03:00
SKT2
2 Intern
•
1.3K Posts
0
April 17th, 2009 03:00
SKT2
2 Intern
•
1.3K Posts
0
April 17th, 2009 05:00
RRR
2 Intern
•
5.7K Posts
0
April 17th, 2009 06:00
Where's everybody else ?
Allen Ward
4 Operator
•
2.1K Posts
1
April 17th, 2009 08:00
Ultimately it is Server's responsibility though.
AranH1
2.2K Posts
0
April 17th, 2009 10:00
I have had enough conversations like that with DBAs and Sys Admins, and now the DBAs and Sys Admins never question me when I tell them what update needs to be applied to their servers.
SKT2
2 Intern
•
1.3K Posts
0
April 17th, 2009 10:00
dynamox
9 Legend
•
20.4K Posts
0
April 17th, 2009 10:00
Allen Ward
4 Operator
•
2.1K Posts
0
April 17th, 2009 13:00
I'm sure it will come up again...
dynamox
9 Legend
•
20.4K Posts
0
April 17th, 2009 16:00
SKT2
2 Intern
•
1.3K Posts
0
April 17th, 2009 18:00
Allen Ward
4 Operator
•
2.1K Posts
0
April 30th, 2009 10:00
Somehow the introduction of SAN storage has shifted the automatic blame for everything from Network to Storage. I highly recommend to every person out there considering a career with Storage: Seriously consider some courses on "Duck, Cover, and Deflect" as well as "Persuading and Influencing". Knowing your stuff is important, but if you can't convince other people of the obvious then you will spend more time defending your environment than you will maintaining and improving it!!!
SKT2
2 Intern
•
1.3K Posts
0
May 8th, 2009 07:00
Allen Ward
4 Operator
•
2.1K Posts
0
May 11th, 2009 12:00
Other shops probably deploy it more widely if they are discovering databases (which we don't do yet) or use the Common Mapping Agent to discover hosts (which we don't do either). Some other people like to have it out there as a diagnostic tool for hosts connectivity issues.
Am I missing anything? Anyone using it for anything else?