Start a Conversation

This post is more than 5 years old

Solved!

Go to Solution

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?

2.2K Posts

April 17th, 2009 08:00

Since PowerPath is an integral component for SAN connectivity it falls under the SAN admins to mange the versions/upgrades. I just have to coordinate with server admins for the upgrades of PowerPath during our one maintenance window a month.

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 ;-)

2 Intern

 • 

5.7K Posts

April 17th, 2009 03:00

It depends per customer, but in most cases it's something the OS team as well as the SAN and Storage teams decide together. Mostly we're not up to date up to the letest release, but 2 versions behind. Except the Linux guys.... those always like the latest and greatest versions.

2 Intern

 • 

1.3K Posts

April 17th, 2009 03:00

I know my question is different. But your comments would be appreciated

2 Intern

 • 

1.3K Posts

April 17th, 2009 05:00

i know it would differ by customer; But i wanted to know know the trend

2 Intern

 • 

5.7K Posts

April 17th, 2009 06:00

Ok, you just got my experience with PP. That's a start.

Where's everybody else ?

4 Operator

 • 

2.1K Posts

April 17th, 2009 08:00

In our shop the Server guys are responsible for PowerPath but the Storage guys provide guidance and support (since we have more experience with it than they do). But we Storage guys used to be Server guys, so there is a bit of a blend.

Ultimately it is Server's responsibility though.

2.2K Posts

April 17th, 2009 10:00

Well put :D

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.

2 Intern

 • 

1.3K Posts

April 17th, 2009 10:00

that reflects a responsible SAN admin..

9 Legend

 • 

20.4K Posts

April 17th, 2009 10:00

in my shop we try to keep as close to supported versions as possible. I encourage my SA people to upgrade whenever they do their regular maintenance. They do complain how much "more" work it is ..blah blah. My answer is simple "Sure ..don't upgrade, when senior management asks me why this box is having issues or we can't get EMC to help us because you are running a 5 year old version of PP, i am going to say that's because your system administrator did not think it was important and didn't care to do it."

4 Operator

 • 

2.1K Posts

April 17th, 2009 13:00

I've got a very similar issue with my Unix admins. They are convinced that PowerPath is too much work and have tried twice now to convince everyone involved that they should just be running MPIO. We've shut them down twice and said that without PowerPath we won't allocate the SAN storage.

I'm sure it will come up again...

9 Legend

 • 

20.4K Posts

April 17th, 2009 16:00

oh yeah ..'cause you know the first time they start having problems with MPIO it will be "SAN" problem.

2 Intern

 • 

1.3K Posts

April 17th, 2009 18:00

thread progress is quite interesting than i expected!

4 Operator

 • 

2.1K Posts

April 30th, 2009 10:00

You know it dynamox!

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!!!

2 Intern

 • 

1.3K Posts

May 8th, 2009 07:00

I would like to know your openion on Solution Enabler (symcli) too, please

4 Operator

 • 

2.1K Posts

May 11th, 2009 12:00

Well, from my perspective I only want SE deployed on hosts that absolutely require it. I know some people like to have it available on all SAN attached (at least Symm attached) hosts, but we only put it where we absolutely require the control functions. So, it goes on the ECC agent hosts that require it, as well as application hosts where they have to control TimeFinder functions. We do not allow it to be installed anywhere else.

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?
No Events found!

Top